Springer-Verlag), pp. 109-118. A Generic Neutral Model for Quantitative Comparison of Genot

Förster resonance energy transferFrom Wikipedia, the free encyclopediaJump to: navigation, searchJablonski diagram of FRET with typical timescales indicatedFörster resonance energy transfer (FRET), fluorescence resonance energy transfer (FRET), resonance energy transfer (RET) orelectronic energy transfer (EET) is a mechanism describing energy transfer between two light-sensitive molecules (chromophores).[1] A donor chromophore, initially in its electronic excited state, may transfer energy to an acceptor chromophore through nonradiative dipole–dipole coupling.[2] The efficiency of this energy transfer is inversely proportional to the sixth power of the distance between donor and acceptor, making FRET extremely sensitive to small changes in distance.[3]Measurements of FRET efficiency can be used to determine if two fluorophores are within a certain distance of each other.[4] Such measurements are used as a research tool in fields including biology and chemistry.FRET is analogous to near-field communication, in that the radius of interaction is much smaller than the wavelength of light emitted. In the near-field region, the excited chromophore emits a virtual photon that is instantly absorbed by a receiving chromophore. These virtual photons are undetectable, since their existence violates the conservation of energy and momentum, and hence FRET is known as a radiationless mechanism. Quantum electrodynamical calculations have been used to determine that radiationless (FRET) and radiative energy transfer are the short- and long-range asymptotes of a single unified mechanism.[5][6]Contents[hide]∙ 1 Terminology∙ 2 Theoretical basis∙ 3 Experimental co nfirmation of the Förster resonance energy transfer theory∙ 4 Methods to measure FRET efficiencyo 4.1 Sensitized emissiono 4.2 Photobleaching FRETo 4.3 Lifetime measurements∙ 5 Fluorophores used for FRETo 5.1 CFP-YFP pairso 5.2 BRETo 5.3 Homo-FRET∙ 6 Applicationso 6.1 Biology∙7 Other methods∙8 See also∙9 References∙10 External linksTerminology[edit]Förster resonance energy transfer is named after the German scientist Theodor Förster.[7] When both chromophores are fluorescent, the term "fluorescence resonance energy transfer" is often used instead, although the energy is not actually transferred by fluorescence.[8][9]In order to avoid an erroneous interpretation of the phenomenon that is always a nonradiative transfer of energy (even when occurring between two fluorescent chromophores), the name "Förster resonance energy transfer" is preferred to "fluorescence resonance energy transfer;" however, the latter enjoys common usage in scientific literature.[10] It should also be noted that FRET is not restricted to fluorescence. It can occur in connection with phosphorescence as well.[8]Theoretical basis[edit]The FRET efficiency () is the quantum yield of the energy transfer transition, i.e. the fraction of energy transfer event occurring per donor excitation event:[11]where is the rate of energy transfer, the radiative decayrate, and the 's are the rate constants of any other de-excitation pathways.[12]The FRET efficiency depends on many physical parameters that can be grouped as follows:∙The distance between the donor and the acceptor (typically in the range of 1-10 nm)∙The spectral overlap of the donor emission spectrum and the acceptor absorption spectrum.∙The relative orientation of the donor emission dipole moment and the acceptor absorption dipole moment.depends on the donor-to-acceptor separation distance with an inverse 6th power law due to the dipole-dipole coupling mechanism:with being the Förster distance of this pair of donor andacceptor, i.e. the distance at which the energy transfer efficiency is 50%.[12]The Förster distance depends on the overlap integral of the donor emission spectrum with the acceptor absorption spectrum and their mutual molecular orientation as expressed by the following equation.[13][14]where is the fluorescence quantum yield of the donor in the absence of the acceptor, κ2 is the dipole orientation factor, is the refractive index of the medium, is Avogadro's number, and is the spectral overlap integral calculated aswhere is the normalized donor emission spectrum, and is the acceptor molar extinction coefficient.[15] The orientation factor κ is given by,Where denotes the normalized transition dipole moment of therespective fluorophore and denotes the normalized inter-fluorophore displacement. κ2 =2/3 is often assumed. This value is obtained when both dyes are freely rotating and can be considered to beisotropically oriented during the excited state lifetime. If either dye is fixed or not free to rotate, then κ2 =2/3 will not be a valid assumption. In most cases, however, even modest reorientation of the dyes results in enough orientational averaging that κ2 = 2/3 doesnot result in a large error in the estimated energy transfer distance due to the sixth power dependence of R0 on κ2. Even when κ2 is quite different from 2/3 the error can be associated with a shift in R0 and thus determinations of changes in relative distance for a particular system are still valid. Fluorescent proteins do not reorient on a timescale that is faster than their fluorescence lifetime. In this case 0 ≤ κ2≤ 4.[15]The FRET efficiency relates to the quantum yield and the fluorescence lifetime of the donor molecule as follows:[16]where and are the donor fluorescence lifetimes in the presenceand absence of an acceptor, respectively, or aswhere and are the donor fluorescence intensities with and without an acceptor, respectively.Experimental confirmation of the Förster resonance energy transfer theory[edit]The inverse sixth-power distance dependence of Förster resonance energy transfer was experimentally confirmed by Wilchek, Edelhoch and Brand[17][18] using tryptophyl peptides. Stryer, Haugland and Yguerabide[19] also experimentally demonstrated the theoretical dependence ofFörster resonance ene rgy transfer on the overlap integral by using a fused indolosteroid as a donor and a ketone as an acceptor. However, a lot of contradictions of special experiments with the theory was oserved. The reason is that the theory has approximate character and gives overstimated distances of 50-100 Angstrems (Vekshin N.L. Energy Transfer in Macromolecules, SPIE, 1997; Vekshin N.L. Photonics of Biopolymers, Springer, 2002).Methods to measure FRET efficiency[edit]In fluorescence microscopy, fluorescence confocal laser scanning microscopy, as well as in molecular biology, FRET is a useful tool to quantify molecular dynamics in biophysics and biochemistry, such as protein-protein interactions, protein–DNA interactions, and protein conformational changes. For monitoring the complex formation between two molecules, one of them is labeled with a donor and the other with an acceptor. The FRET efficiency is measured and used to identify interactions between the labeled complexes. There are several ways of measuring the FRET efficiency by monitoring changes in the fluorescence emitted by the donor or the acceptor.[20]Sensitized emission[edit]One method of measuring FRET efficiency is to measure the variationin acceptor emission intensity.[14] When the donor and acceptor are in proximity (1–10 nm) due to the interaction of the two molecules, the acceptor emission will increase because of the intermolecularFRET from the donor to the acceptor. For monitoring protein conformational changes, the target protein is labeled with a donor and an acceptor at two loci. When a twist or bend of the protein brings the change in the distance or relative orientation of the donor and acceptor, FRET change is observed. If a molecular interaction or a protein conformational change is dependent on ligand binding, this FRET technique is applicable to fluorescent indicators for the ligand detection.Photobleaching FRET[edit]FRET efficiencies can also be inferred from the photobleaching rates of the donor in the presence and absence of an acceptor.[14] This method can be performed on most fluorescence microscopes; one simply shines the excitation light (of a frequency that will excite the donor but not the acceptor significantly) on specimens with and without the acceptor fluorophore and monitors the donor fluorescence (typically separated from acceptor fluorescence using a bandpass filter) over time. The timescale is that of photobleaching, which is seconds to minutes, with fluorescence in each curve being given bywhere is the photobleaching decay time constant and depends on whether the acceptor is present or not. Since photobleaching consists in the permanent inactivation of excited fluorophores, resonance energy transfer from an excited donor to an acceptor fluorophore prevents the photobleaching of that donor fluorophore, and thus high FRET efficiency leads to a longer photobleaching decay time constant:where and are the photobleaching decay time constants of thedonor in the presence and in the absence of the acceptor, respectively. (Notice that the fraction is the reciprocal of that used for lifetime measurements).This technique was introduced by Jovin in 1989.[21] Its use of anentire curve of points to extract the time constants can give it accuracy advantages over the other methods. Also, the fact that time measurements are over seconds rather than nanoseconds makes it easierthan fluorescence lifetime measurements, and because photobleaching decay rates do not generally depend on donor concentration (unless acceptor saturation is an issue), the careful control of concentrations needed for intensity measurements is not needed. It is, however, important to keep the illumination the same for the with- and without-acceptor measurements, as photobleaching increases markedly with more intense incident light.Lifetime measurements[edit]FRET efficiency can also be determined from the change in the fluorescence lifetime of the donor.[14] The lifetime of the donor will decrease in the presence of the acceptor. Lifetime measurements of FRET are used in Fluorescence-lifetime imaging microscopy.Fluorophores used for FRET[edit]If the linker is intact, excitation at the absorbance wavelength of CFP (414nm) causes emission by YFP (525nm) due to FRET. If the linker is cleaved by a protease, FRET is abolished and emission is at the CFP wavelength (475nm).CFP-YFP pairs[edit]One common pair fluorophores for biological use is a cyan fluorescent protein (CFP) – yellow fluorescent protein (YFP) pair.[22] Both are color variants of green fluorescent protein (GFP). Labeling with organic fluorescent dyes requires purification, chemical modification, and intracellular injection of a host protein. GFP variants can be attached to a host protein by genetic engineering which can be more convenient. Additionally, a fusion of CFP and YFP linked by a protease cleavage sequence can be used as a cleavage assay.[23]BRET[edit]A limitation of FRET is the requirement for external illumination to initiate the fluorescence transfer, which can lead to background noise in the results from direct excitation of the acceptor or to photobleaching. To avoid this drawback, Bioluminescence Resonance Energy Transfer (or BRET) has been developed.[24] This technique uses a bioluminescent luciferase (typically the luciferase from Renilla reniformis) rather than CFP to produce an initial photon emission compatible with YFP.Homo-FRET[edit]In general, "FRET" refers to situations where the donor and acceptor proteins (or "fluorophores") are of two different types. In many biological situations, however, researchers might need to examine the interactions between two, or more, proteins of the same type—or indeed the same protein with itself, for example if the protein folds or forms part of a polymer chain of proteins[25] or for other questions of quantification in biological cells.[26]Obviously, spectral differences will not be the tool used to detect and measure FRET, as both the acceptor and donor protein emit light with the same wavelengths. Yet researchers can detect differences in the polarisation between the light which excites the fluorophores andthe light which is emitted, in a technique called FRET anisotropy imaging; the level of quantified anisotropy (difference in polarisation between the excitation and emission beams) then becomes an indicative guide to how many FRET events have happened.[27]Applications[edit]Biology[edit]FRET has been used to measure distance and detect molecular interactions in a number of systems and has applications in biology and chemistry.[28] FRET can be used to measure distances between domains in a single protein and therefore to provide information about protein conformation.[29] FRET can also detect interaction between proteins.[30] Applied in vivo, FRET has been used to detect the location and interactions of genes and cellular structures including intergrins and membrane proteins.[31] FRET can be used to obtain information about metabolic or signaling pathways.[32] FRET is also used to study lipid rafts in cell membranes.[33]FRET and BRET are also the common tools in the study of biochemical reaction kinetics and molecular motors.The applications of Fluorescence Resonance Energy Transfer (FRET) have expanded tremendously in the last 25 years, and the technique has become a staple technique in many biological and biophysical fields. FRET can be used as spectroscopic ruler in various areas such as structural elucidation of biological molecules and their interactions in vitro assays, in vivo monitoring in cellular research, nucleic acid analysis, signal transduction, light harvesting and metallic nanomaterial etc. Based on the mechanism of FRET a variety of novel chemical sensors and biosensors have been developed.[34]Other methods[edit]A different, but related, mechanism is Dexter Electron Transfer.An alternative method to detecting protein–protein proximity is the bimolecular fluorescence complementation (BiFC) where two halves of a YFP are fused to a protein. When these two halves meet they form a fluorophore after about 60 s – 1 hr.[35]See also[edit]∙Förster coupling∙Surface energy transfer∙Dexter electron transfer∙Time-resolved fluorescence energy transferReferences[edit]1.Jump up ^ Cheng, Ping-Chin (2006). "The Contrast Formation in OpticalMicroscopy". In Pawley, James B. Handbook Of Biological Confocal Microscopy(3rd ed.). New York, NY: Springer. pp. 162–206. doi:10.1007/978-0-387-45524-2_8. ISBN 978-0-387-25921-5.2.Jump up ^ Helms, Volkhard (2008). "Fluorescence Resonance Energy Transfer".Principles of Computational Cell Biology. Weinheim: Wiley-VCH. p. 202.ISBN 978-3-527-31555-0.3.Jump up ^ Harris, Daniel C. (2010). "Applications of Spectrophotometry".Quantitative Chemical Analysis (8th ed.). New York: W. H. Freeman and Co.pp. 419–44. ISBN 978-1-4292-1815-3.4.Jump up ^ Zheng, Jie (2006). "Spectroscopy-Based Quantitative FluorescenceResonance Energy Transfer Analysis". In Stockand, James D.; Shapiro, MarkS. Ion Channels: Methods and Protocols. Methods in Molecular Biology,Volume 337. Totowa, NJ: Humana Press. pp. 65–77. doi:10.1385/1-59745-095-2:65. ISBN 978-1-59745-095-9.5.Jump up ^ Andrews, David L. (1989). "A unified theory of radiative andradiationless molecular energy transfer". Chemical Physics135 (2): 195–201. Bibcode:1989CP....135..195A. doi:10.1016/0301-0104(89)87019-3.6.Jump up ^ Andrews, David L; Bradshaw, David S (2004). "Virtual photons,dipole fields and energy transfer: A quantum electrodynamical approach".European Journal of Physics25 (6): 845. doi:10.1088/0143-0807/25/6/017.7.Jump up ^Förster, Theodor (1948). "Zwischenmolekulare Energiewanderung undFluoreszenz" [Intermolecular energy migration and fluorescence]. Annalender Physik (in German) 437: 55–75. Bibcode:1948AnP...437...55F.doi:10.1002/andp.19484370105.8.^ Jump up to: a b Valeur, Bernard; Berberan-Santos, Mario (2012). "ExcitationEnergy Transfer". Molecular Fluorescence: Principles and Applications, 2nded. Weinheim: Wiley-VCH. pp. 213–261. doi:10.1002/9783527650002.ch8.ISBN 9783527328376.9.Jump up ^ FRET microscopy tutorial from Olympus10.Jump up ^Glossary of Terms Used in Photochemistry (3rd ed.). IUPAC. 2007.p. 340.11.Jump up ^ Moens, Pierre. "Fluorescence Resonance Energy Transferspectroscopy". Retrieved July 14, 2012.12.^ Jump up to: a b Schaufele, Fred; Demarco, Ignacio; Day, Richard N. (2005)."FRET Imaging in the Wide-Field Microscope". In Periasamy, Ammasi; Day, Richard. Molecular Imaging: FRET Microscopy and Spectroscopy. Oxford:Oxford University Press. pp. 72–94. doi:10.1016/B978-019517720-6.50013-4.ISBN 978-0-19-517720-6.13.Jump up ^Förster, Th. (1965). "Delocalized Excitation and ExcitationTransfer". In Sinanoglu, Oktay. Modern Quantum Chemistry. IstanbulLectures. Part III: Action of Light and Organic Crystals3. New York and London: Academic Press. pp. 93–137. Retrieved 2011-06-22.14.^ Jump up to: a b c d Clegg, Robert (2009). "Förster resonance energytransfer—FRET: what is it, why do it, and how it's done". In Gadella,Theodorus W. J. FRET and FLIM Techniques. Laboratory Techniques inBiochemistry and Molecular Biology, Volume 33. Elsevier. pp. 1–57.doi:10.1016/S0075-7535(08)00001-6. ISBN 978-0-08-054958-3.15.^ Jump up to: a b Demchenko, Alexander P. (2008). "Fluorescence DetectionTechniques". Introduction to Fluorescence Sensing. Dordrecht: Springer.pp. 65–118. doi:10.1007/978-1-4020-9003-5_3. ISBN 978-1-4020-9002-8. 16.Jump up ^ Majoul, Irina; Jia, Yiwei; Duden, Rainer (2006). "PracticalFluorescence Resonance Energy Transfer or Molecular Nanobioscopy of Living Cells". In Pawley, James B. Handbook Of Biological Confocal Microscopy (3rd ed.). New York, NY: Springer. pp. 788–808. doi:10.1007/978-0-387-45524-2_45. ISBN 978-0-387-25921-5.17.Jump up ^ Template:ISRAEL JOURNAL OF CHEMISTRY. Vol. 1. No. Sa. 196318.Jump up ^ {Edelhoch, H., Brand, L., Wilchek, M. (1967). "Fluorescencestudies with tryptophyl peptides". Biochemistry 6 (2): 547–559.doi:10.1021/bi00854a024. PMID 6047638}19.Jump up ^ Lakowicz, Joseph R., ed. (1991). Principles. New York: PlenumPress. p. 172. ISBN 978-0-306-43875-2.20.Jump up ^ "Fluorescence Resonance Energy Transfer Protocol". WellcomeTrust. Retrieved 24 June 2012.[dead link]21.Jump up ^Szöllősi, János; Alexander, Denis R. (2007). "The Application ofFluorescence Resonance Energy Transfer to the Investigation ofPhosphatases". In Klumpp, Susanne; Krieglstein, Josef. ProteinPhosphatases. Methods in Enzymology, Volume 366. Amsterdam: Elsevier.pp. 203–24. doi:10.1016/S0076-6879(03)66017-9. ISBN 978-0-12-182269-9. 22.Jump up ^ Periasamy, Ammasi (July 2001). "Fluorescence resonance energytransfer microscopy: a mini review". Journal of Biomedical Optics6 (3): 287–291. Bibcode:2001JBO.....6..287P. doi:10.1117/1.1383063.PMID 11516318.23.Jump up ^ Nguyen, AW; Daugherty, PS (March 2005). "Evolutionaryoptimization of fluorescent proteins for intracellular FRET.". Naturebiotechnology23 (3): 355–60. doi:10.1038/nbt1066. PMID 15696158.24.Jump up ^ Bevan, Nicola; Rees, Stephen (2006). "Pharmaceutical Applicationsof GFP and RCFP". In Chalfie, Martin; Kain, Steven R. Green FluorescentProtein: Properties, Applications and Protocols. Methods of Biochemical Analysis, Volume 47 (2nd ed.). Hoboken, NJ: John Wiley & Sons. pp. 361–90. doi:10.1002/0471739499.ch16. ISBN 978-0-471-73682-0.25.Jump up ^ Gautier, I.; Tramier, M.; Durieux, C.; Coppey, J.; Pansu, R.B.;Nicolas, J.-C.; Kemnitz, K.; Coppey-Moisan, M. (2001). "Homo-FRETMicroscopy in Living Cells to Measure Monomer-Dimer Transition of GFP-Tagged Proteins". Biophysical Journal80 (6): 3000–8.Bibcode:2001BpJ....80.3000G. doi:10.1016/S0006-3495(01)76265-0.PMC 1301483. PMID 11371472.26.Jump up ^ Bader, Arjen N.; Hofman, Erik G.; Voortman, Jarno; Van Bergen EnHenegouwen, Paul M.P.; Gerritsen, Hans C. (2009). "Homo-FRET ImagingEnables Quantification of Protein Cluster Sizes with SubcellularResolution". Biophysical Journal97 (9): 2613–22.Bibcode:2009BpJ....97.2613B. doi:10.1016/j.bpj.2009.07.059. PMC 2770629.PMID 19883605.27.Jump up ^ Gradinaru, Claudiu C.; Marushchak, Denys O.; Samim, Masood;Krull, Ulrich J. (2010). "Fluorescence anisotropy: From single molecules to live cells". The Analyst135 (3): 452–9. Bibcode:2010Ana...135..452G.doi:10.1039/b920242k. PMID 20174695.28.Jump up ^ Lakowicz, Joseph R. (1999). Principles of fluorescencespectroscopy (2nd ed.). New York, NY: Kluwer Acad./Plenum Publ. pp. 374–443. ISBN 978-0-306-46093-7.29.Jump up ^ Truong, Kevin; Ikura, Mitsuhiko (2001). "The use of FRET imagingmicroscopy to detect protein–protein interactions and proteinconformational changes in vivo". Current Opinion in Structural Biology11(5): 573–8. doi:10.1016/S0959-440X(00)00249-9. PMID 11785758.30.Jump up ^ Pollok, B; Heim, R (1999). "Using GFP in FRET-basedapplications". Trends in Cell Biology9 (2): 57–60. doi:10.1016/S0962-8924(98)01434-2. PMID 10087619.31.Jump up ^ Sekar, R. B.; Periasamy, A (2003). "Fluorescence resonance energytransfer (FRET) microscopy imaging of live cell protein localizations". The Journal of Cell Biology160 (5): 629–33. doi:10.1083/jcb.200210140.PMC 2173363. PMID 12615908.32.Jump up ^ Ni, Qiang; Zhang, Jin (2010). "Dynamic Visualization of CellularSignaling". In Endo, Isao; Nagamune, Teruyuki. Nano/Micro Biotechnology.Advances in Biochemical Engineering/Biotechnology, Volume 119. Springer.pp. 79–97. Bibcode:2010nmb..book...79N. doi:10.1007/10_2008_48.ISBN 978-3-642-14946-7. PMID 19499207.33.Jump up ^ Silvius, John R.; Nabi, Ivan Robert (2006). "Fluorescence-quenching and resonance energy transfer studies of lipid microdomains in model and biological membranes (Review)". Molecular Membrane Biology23(1): 5–16. doi:10.1080/09687860500473002. PMID 16611577.34.Jump up ^ S. A., Hussain et al. (2015). "Fluorescence Resonance EnergyTransfer (FRET) sensor" (PDF). J. Spectrosc. Dyn.5 (7): 1–16.35.Jump up ^ Hu, Chang-Deng; Chinenov, Yurii; Kerppola, Tom K. (2002)."Visualization of Interactions among bZIP and Rel Family Proteins in Living Cells Using Bimolecular Fluorescence Complementation". Molecular Cell9(4): 789–98. doi:10.1016/S1097-2765(02)00496-3. PMID 11983170。

ORIGINAL PAPERAffordance-based design methods for innovative design,redesign and reverse engineeringJonathan R.A.Maier ÆGeorges M.FadelReceived:13December 2008/Accepted:5January 2009/Published online:12March 2009ÓSpringer-Verlag London Limited 2009Abstract Rather than developing methods to address problems as they occur,the effort in this paper is to formulate methods based on an explicit theory.Methods developed in this way have more scientific rigor because underlying propositions and assumptions are clearly articulated,thus the applicability and limitations of the methods are well defined.The underlying theory used in this work is that of affordance-based design,which has been developed by the authors in a recent series of papers,and is based in turn on the theory of affordances from perceptual psychology.This paper extends affordance-based design into prescriptive methods.A broad affordance-based design process is introduced together with methods for documenting affordances,methods for design-ing individual affordances,an affordance-based method for reverse engineering and redesign,the affordance structure matrix,and affordance-based selection matrices.Engineer-ing examples used to illustrate the methods include the high level design of an automobile,comfort to automobile passengers,the meshing of gears,wear of gears,a vacuum cleaner,and automotive window switches.Keywords Affordances ÁAffordance-based design ÁDesign methodology1IntroductionThe existing state of the art in design science is character-ized by much stronger methods than theory.Some methods,such as those of axiomatic design are based on an explicit theory (Suh 1990,2001).Other methods,such as Quality Function Deployment (Clausing 1994)are empirical in nature,derived from what appears to work in practice,rather than any underlying theory.Similarly,advances in design-for-x methods are typically ad hoc in nature,developed only in response to a particular need ‘‘for x’’.Herbert Simon,in criticizing the lack of theoretical basis behind design methods in the 1960s described such ad hoc methods as ‘‘cook-booky’’(Simon 1996).This is in con-trast to the methods used in other engineering sciences,and even mathematics,where methods to solve a particular problem,say of heat transfer,are based on an underlying theory,for example,Newton’s Law of Cooling.We pro-pose that more rigorous methods in design should likewise be based on theory first,and then validated and refined by their use in industry.The effort in this paper is therefore to present a series of practical methods based on the theory of affordance-based design elaborated in previous work by the authors (Maier and Fadel 2001,2002,2003,2005a ,b ,2008).In Sect.2,the fundamentals of affordance-based design are summa-rized and a high level method for affordance-based design is presented,describing the steps needed to execute an overall design project from inception to finished product.Given the variability inherent in different design projects,the steps described are of necessity somewhat general.More specific methods for documenting the affordances pertaining to a design project are offered in Sect.3.A method for designing individual affordances is given in Sect.4,and a method for reverse engineering and redesign is presented in Sect.5.A more detailed tool for reverse engineering and concept exploration,the affordance structure matrix,is discussed in Sect.6,followed by similar affordance-based selection matrices in Sect.7.J.R.A.Maier ÁG.M.Fadel (&)Clemson Research in Engineering Design and Optimization (CREDO)Laboratory,Department of Mechanical Engineering,Clemson University,Clemson,SC 29634-0921,USA e-mail:fgeorge@Res Eng Design (2009)20:225–239DOI 10.1007/s00163-009-0064-7Closing remarks and suggestions for future work are given in Sect.8.2A high level method for affordance-based designIn previous work,the authors(Maier and Fadel2001,2002, 2003,2005a,b,2008)discuss several theoretical aspects of affordance-based design that can now be used to construct prescriptive methods.Whereas the concept of affordance was originally introduced in thefield of perceptual psy-chology(Gibson1979),in the context of engineering design,we define an affordance as a relationship between two subsystems in which a potential behavior can occur that would not be possible with either subsystem in isola-tion(Maier and Fadel2008).Such relationships are particularly appropriate in describing the potential inter-action between human users and artifacts,which are called artifact–user affordances(AUA).However,affordances can also describe potential behaviors between two artifact subsystems,which are then termed artifact–artifact affor-dances(AAA).An example of an AUA is a vehicle affording transportation to its driver and occupants.An example of an AAA is two gears affording power trans-mission.In design,AAA usually exist at a lower level to serve higher level AUA.Important properties of affordances include:•Complementarity:an affordance describes an interac-tion between two subsystems.The affordance is relative to both subsystems and cannot exist with respect to either subsystem in isolation.For example,a single gear cannot transmit power.•Polarity:affordances can be either positive or negative, depending upon whether the potential behavior has beneficial or harmful consequences.A vehicle afford-ing transportation to an occupant is a positive affordance,whereas that vehicle affording injury to a pedestrian is a negative affordance.•Multiplicity:systems can have multiple affordances,as in the vehicle example discussed above.•Quality:affordances can be of varying quality depend-ing upon how well the subsystems support the potential behavior.The affordance of occupant comfort is of higher quality in a luxury car,for example,than it is in an economy car.•Form dependence:affordances depend on the physical structures of artifacts,unlike functions which are form independent.The design process can be viewed in terms of affor-dances,in light of the concept of affordance,as discussed previously by the authors(Maier and Fadel2003;Maier et al.2008a,b).The impetus for any design project can be understood in terms of creating and changing affordances. The design process can be viewed simply as the specifi-cation of an artifact that possesses certain desired affordances,and does not possess certain undesired affor-dances.An artifact with more(higher quality)positive affordances is considered better,while an artifact with more negative(or lower quality)affordances is considered worse.Readers may consult the authors’previous discus-sion of these ideas for a more in-depth treatment. Meanwhile,taking all these precepts into consideration,a high level affordance-based design process is depicted in Fig.1.The affordance-based design process begins with moti-vation:a perceived market need,a novel idea,a scheduled product redesign,etc.The parent company,if there is one, would typically provide this information and initial product cost targets,schedule targets,marketing targets,etc.,to the designer or design team.In affordance-based design,thefirst task for the designer(s),is to determine the AUA that the artifact should have and not have.Because of polarity,the designers should identify both positive affordances and negative affordances(an aspect which is unique to affor-dance-based design).And because of complementarity,the affordances will depend on different users,so the design-er(s)must identify the different users,grouping them as convenient,and then interviewing various users to deter-mine wanted and unwanted affordances.Following the methods presented below in Sect.3for creating affordance structures,the affordances should then be prioritized(the highest priority affordance can also be considered the design drivers);andfinally,one or more affordance struc-tures can be constructed.As described in Sect.3,the generic affordance structure template should also be con-sulted in the process.The second task for designers is to ideate to generate concepts for the artifact’s overall architecture and com-ponents.Various established ideation methods can be used here,such as brainstorming,TRIZ(cf.,Altshuller1984, 1996,1997,2000),patent searches,Synectics(cf.,Rourkes 1988),IDEO’s‘‘deep dive’’process(cf.,Kelly and Littman 2001),etc.External references may also be consulted for ideas,such as the Internet,traditional libraries,industry catalogues,etc.Sketches of each concept are typically produced using these ideation methods;however,these sketches are particularly important for affordance-based design,because of form dependence.Each concept should be analyzed for how well it satisfies the desired positive affordances with reference back to the affordances docu-mented in the affordance structure(s)created in the previous step.The third task for designers is to analyze and refine the affordances of the concepts generated in the previous stage.This involves modifying the characteristics of concepts in order to modify their affordances,as well as analyzing the negative affordances of each concept,and modifying their characteristics accordingly to remove those negative affordances.A helpful tool for this step of affordance-based design is the Affordance Structure Matrix as discussed in Sect.6.Concepts from various architecture concepts can also be combined,switched around,and refined in order to modify the affordances of the overall artifact.The construction of prototypes of concept archi-tectures or components may also be needed in order to better understand the affordances of each concept,again because of form dependence.The fourth task for designers is to select a preferred architecture.Various selection methods can be used in this process,including the Gallery method (Hellfritz 1978),Pugh decision matrices (Pugh 1996,p.167–176),a Selection Decision Support Problem (Kuppuraju et al.1985),Utility theory (cf.,Hazelrigg 1996),or the affor-dance selection matrices discussed in Sect.7.However,the decision should ultimately rest on how well each concept satisfies the desired positive affordances while eliminating or minimizing undesired negative affordances,giving preference to concepts with higher quality affordances,and preference to concepts with extra positive affordances.Note that the theory of affordances does not suggest a preferred selection method,but it does inform the criteria to be used in the decision-making process.The fifth task is to determine the artifact–artifact affor-dances (AAA)that should exist between the subsystems in the preferred architecture.For example,the transmission of forces,heat,fluids,electricity,and informationbetweenFig.1Overview of theaffordance-based design processsubsystems must be afforded.As these individual AAAs are elucidated,they should be added to the affordance structure(s)previously created.The sixth task is to design individual affordances.Accordingly,in Sect.4,methods are described for designing positive AUA,and positive AAA,as well as methods for designing against negative AUA and negative AAA.These methods must be executed on the preferred concept architecture and components in order to specify detailed artifact characteristics such that the detailed arti-fact does indeed posses all the desired positive affordances and does not possess any of (or at least minimizes)the undesired negative affordances,as articulated in the affordance structure.3A method for creating affordance structures One of the first steps in popular function-based system-atic design methods and their derivatives is the creation of some type of function structure,a graphical repre-sentation of required functions and some relationships between them.After briefly reviewing the two most prevalent types of function structures,a method is then introduced for creating four different types of affordance structures.The purpose of a function structure is to organize the functions of an artifact in a coherent manner.One type of function structure called a function tree organizes functions hierarchically,from a top-level general description of the product function down to low level ‘‘atomic’’functions that can be decomposed no further.An example function tree is shown in Fig.2.Another type of function structure organizes functions in terms of flows;specifically,flows of energy,matter,and information through and between product functions.An example of this type of function structure is shown in Fig.3.This type of function structure can,but often does not,capture the hierarchy that a function tree captures.3.1Affordance structures versus function structures The purpose of an affordance structure is similar to that of a function structure in that,whereas a function structure organizes functions,an affordance structure organizes affordances.However,because of the different properties of functions and affordances,affordance structures cannot be organized in the same ways as function structures.Recall that a crucial difference between functions and affordances is that functions are form independent,whereas affordances are form dependent .The idea behind system-atic function-based design methods [such as Pahl and Beitz (1996)]is that customer demands and wishes can be translated into functions,which can then be accomplished by physical working principles,which are then embodied by particular physical parts.It seems reasonable that fol-lowing the rigorous steps of these methods should produce more carefully considered designs;however,it is not clear whether this is a consequence of the methods being func-tional as opposed to merely systematic.In fact,there is at least one noted example,coming from a very successful design firm,that a systematic method,which does not explicitly consider functions but which does move very quickly to form solutions is creative,effective,and fast (Kelly and Littman 2001).In this work,therefore,a sys-tematic approach is maintained;however,the concept of affordance is used as the fundamental theoretical construct,rather than the concept of function.3.2A method to create specification level affordancestructures for original design 3.2.1Step 1:understanding,gathering,and expressinguser needs in terms of affordances The first step in building an affordance structure is to express the users demands and wishes as affordances.From a psychological point of view,this should not be difficult since ecological psychology predicts that people perceive their environment (including artifacts)already in termsofFig.2Example function tree (Shupe et al.1987)Fig.3Example function flow diagram (Pahl and Beitz 1996)affordances(Gibson1979).However,generally it is easier to express affordances in terms of an affordance relation-ship rather than a one word name for the affordance.For example,it would usually be more appropriate to say that‘a vehicle must afford transportation’,than merely to list the affordance of‘transportability.’Note that users include not just end users,but everyone who will interact with the artifact throughout its lifecycle,including designers, manufacturers,marketers,salesmen,end users,mainte-nance personnel,recyclers,trash haulers,etc.At this point,the affordances should be grouped into two categories:positive affordances(what the artifact should afford)and negative affordances(what the artifact should not afford).The output of this step is an unordered list of annotated positive affordances and an unordered list of annotated negative affordances.The annotations should document from which users each affordance was suggested and to which user each affordance pertains.As an example artifact in this section,consider a typical passenger car.The affordances of such a vehicle will be analyzed throughout this section.However,the focus of this paper is on the methods and not on any of the examples,per se,so the vehicle example is pursued at a fairly high level.A short list of positive and negative affordances for the vehicle is shown in Table1.3.2.2Step2:apply generic affordance structure template The second step is to compare the list of affordances with the generic affordance structure template(Fig.4).For affordances that are present in the generic affordance structure template,but do not show-up in the user affor-dance list,the designer must decide whether those affordances are indeed important(even though they did not appear in the user affordances),or whether they can be neglected for this particular project.The generic affordance structure template thus serves two purposes.Thefirst is to guide designers as to what affordances in general they should expect to provide.The second is to catch and include affordances that might otherwise be missed.In the passenger car example,there are many affordances in the template that do not appear in Table1.These additional affordances pertaining,for example,to aesthetics,manu-facturing,maintenance and the like,should be added to the list of artifact affordances.3.2.3Step3:prioritizing affordancesThe third step toward building an affordance structure is to prioritize the affordances.Note that,as a consequence of Arrow’s theorem(cf.,Franssen2005;Scott and Antonsson 1999),this kind of ranking cannot objectively group user’s preferences.Therefore,ultimately the prioritization of affordances really only reflects the designer’s preferences, but this should be informed,as much as possible,by user information.Note that several affordances may share the same priority level.The output of this step is a(tentatively) complete list of prioritized annotated affordances.One prioritization of the affordances for the passenger car is shown in Table2.As indicated in the high level design procedure presented in Sect.2,the resulting design depends critically on the identification of pertinent affor-dances,the subjective assignment of priorities to the individual affordances,the understanding developed of each affordance,and external factors such as time and money available for the design project,market conditions, legal regulations,etc.3.2.4Step4:organizing the affordances into a structure By priority The fourth step is to organize the list of affordances into an ordered structure.This can be done in several ways.One way to organize the affordance structure is hierarchically by priority level.This organization is well-suited to projects where most of the affordances are not those shown in the generic affordance structure,and/or projects for which there are many priority levels.This type of orga-nization also leaves plenty of space to show annotations. Often the most convenient way to represent this type of organization is in the form of a table,as in Table2below.By topic Another way to organize the affordances is by topic,i.e.,how the generic affordance structure is orga-nized.This organization is well-suited to projects where most of the affordances for the project are simply those in the generic affordance structure,and/or projects for which there are few meaningful priority levels.This type of organization can be depicted graphically as a conventional inverted tree,or as a sunshine type of tree as is the generic affordance structure template.In either case,the priority level for each affordance may be shown.However,one drawback to this type of organization is that there is typi-cally not enough space to include all the annotations unless large format paper is used(which is sometimes done for other organized graphs,such as House of Quality matrices).Table1Affordances for a vehicle(annotations omitted)Positive affordances: (The artifact must afford…)Negative affordances:(The artifact must NOT afford…)Transportation of occupants Injuring occupants Transportation of cargo Injuring othersComfort to occupants Damaging other vehicles Entertainment of occupants Pollution to the environment Communication to othersAn example of a topically organized affordance structure is shown in Fig.5.By user group Lastly,an affordance structure may also be organized by user groups to which the affordances pertain.This organization is well-suited to projects where different groups of users have distinct and/or conflicting interests.The annotations contain the necessary information to organize the affordances this way.A drawback to this organization is that the same affordance can appear in multiple user groups.This can lead to a redundant structure if the user groups’interests are relatively homogenous.However,it is easy to indicate the priority level associated with each affordance.An example affordance structure of this type is shown in Table 3.3.2.5Additional remarksIt is a relatively easy and quick process to draw an affor-dance structure organized in any of these ways,so for any given design project it is good practice to prepare multiple affordance structures to see which kind appears most useful for that particular project.Because the affordance struc-tures described above are created at the specification level,and because the specifications are usually given or inter-preted as AUAs,a specification level affordance structure typically contains mostly AUAs and relatively few AAAs.An affordance structure at this point,therefore,is neces-sarily incomplete until after a preferred architecture has been chosen,and the AAAs of that architecture have been determined.Then those AAAs can be added to the original specification level affordance structure to complete a detail level affordancestructure.Fig.4Generic affordance structure templateTable 2Prioritized affordances for a ladder (annotations omitted)Priority Positive affordances:(The artifact must afford …)Negative affordances:(The artifact must NOT afford …)1Transportation of occupants Transportation of cargo Manufacture to workersInjuring occupants Injuring others2Comfort to occupants Aesthetics to buyers and occupantsDamaging other vehicles 3Communication to others Improvement to owners and occupantsMaintenance to owner and workersPollution to the environment Degradation of itself Frustration to occupants 4Entertainment of occupants Retirement to workersSustainability to the environment4A method for designing individual affordances 4.1About designing affordancesSince an affordance is a relationship between two inter-acting subsystems such that a behavior is possible between the two that is not possible with only one of the systems in isolation,in order to design an affordance,we must con-sider pertinent properties of the two different subsystems.In the case of AUA,we usually do not have control over the properties of the user,but we do have control over the properties of the artifact under design.In the case of AAA,we usually have control over both artificial subsystems.Recalling the notion of polarity ,which divides affor-dances into those that are positive,i.e.,beneficial,and those that are negative,i.e.,harmful,then AUA can be positive (?)or negative (-),and AAA can be either positive (?)or negative (-),this yields four major cate-gories of affordances to design:?AUA,-AUA,?AAA,and -AAA.The general method for designing any affor-dance is presented in Fig.6.4.2An example of designing a positive AUAA positive AUA from the passenger car example is that the car must afford comfort to the driver and other fort is an affordance that must be maintained across many vehicle subsystems.For the purposes of illustration,consider only the comfort of the driver seat.Like any other affordance,comfort has the property of varying quality .Furthermore,because of complementarity as discussed above,in order to create a desired affordance,in this case,comfort,the designer must understand the particular characteristics necessary on behalf of both the user and artifact.Following the procedure in Fig.6,we now need to define and to understand what this affordance is.First,we note that this is a positive affordance,something we want the artifact to afford.Second,we note that the relevant users are drivers of the vehicle.These could be further divided into driver,front passenger,and rear passengers.The next steps involve information fortisFig.5Topical affordance structure for passenger car example (annotations omitted).White boxes are shared with the generic affordance structure template.Grey boxes are specific to this exampleTable 3Affordance structure organized by user group (annotations omitted)User groupPositive Affordances:(The artifact must afford …)Negative Affordances:(The artifact must NOT afford …)Occupants 1.Transportation1.Transportation of their cargo2.Aesthetics fort3.Improvement4.Entertainment1.Injuring 3.FrustrationOwner 3.Improvement 3.Maintenance 3.Degradation of their vehicleBuyers 2.Aesthetics Othersmunication4.Sustainability 1.Injuring2.Damaging their vehicles3.PollutionWorkers1.Manufacturing 3.Maintenance 4.Retirementdetermined fundamentally by the structure of the artifact,taken with respect to the user.Fortunately,the physical characteristics of seated automobile drivers have been well studied,and are available in the form of anthropomorphic data (cf.,Webb Associates 1978).The challenge for the designer is to match seat structure to human driver struc-ture.Seat designers may also avail themselves of the many existing examples of seat comfort already in existence,from household and office furniture to existing vehicle seats.The specific properties that determine comfort all need to be identified.These result in this case in a list of relevant anthropomorphic dimensions and seat dimensions.The next step in designing the affordance entails iden-tifying bounds and targets for each property identified above.Since we are dealing with an AUA,we do not have direct control over the properties of the user,but we do have complete control over the properties of the artifact.A principal goal of the design effort is determining the rele-vant bounds and targets for the range of driver dimensions deemed appropriate.A seat for a large SUV or truckneedsFig.6Procedure for designing individual affordancesto accommodate larger size drivers than a compact car,for example.This decision will impact the target height and width of the seat back,elasticity of the seat components, etc.Experimental studies are an invaluable tool in quanti-fying such dimensions.After bounds and ideal targets are identified for seat comfort,the artifact’s affordance structure needs to be consulted to identify which properties that impact the affordance of comfort also impact other positive and neg-ative affordances,for example,pertaining to strength of the seat,or behavior in a crash.What is ideal for creating comfort(such as a very soft seat)may not be strong enough in critical loading conditions,for example.They key here is to select property settings that maximize the affordance of comfort subject to the constraints of maintaining previ-ously designed affordances while attempting to safeguard against negative affordances.As additional affordances are designed,the designers must be vigilant to ensure that previous affordances are indeed maintained,while attempting to identify any new negative affordances.For example,in the vehicle seat,a seat that is too comfortable may in fact promote the driver sleeping,which is a behavior to be designed against!4.3An example of designing against a negative AUA Now that we understand how to create a positive AUA, we need to consider the inverse problem,how to ensure that a negative AUA is not present.Once again,this involves understanding the affordance itself,and then the characteristics of both the artifact and user that bear upon that affordance.The goal in designing against a negative affordance is to ensure that the afforded behavior will not happen.Therefore,instead of specify-ing artifact characteristics to lie within certain ranges, and meeting certain targets,the designer must specify the range that certain artifact characteristics cannot fall within.Recall that physical properties,i.e.,the artifact’s char-acteristics,are not affordances in and of themselves,but they do determine what affordances exist with reference to user characteristics.Therefore,the characteristics that affect positive AUA are the same characteristics that affect negative AUAs.The artifact,in other words,only has one set of characteristics,and that is all the designer has control over.Therefore,when a designer identifies a range in which a characteristic is not allowed to fall in order to avoid a negative affordance,in reality all this does is constrain the target bounds for that same characteristic with respect to the positive affordances that are desired.These ideas are best illustrated by another example.Consider the vehicle seat from the previous section.As discussed above,a negative AUA for the seat is conduciveness for sleep,or ‘sleepability’.After defining the target affordance,the next step,as before,is understanding that affordance.In particular,the designer is interested in identifying what seat configura-tions are especially conducive to sleep.Fortunately,again the designer should utilize existing research into sleepiness in vehicles(cf.,Kanstrup and Lundin2006).This research will guide them toward identifying what ranges of seat properties values are to be avoided,limiting,in effect, design freedom as far as property settings for the earlier designed affordance of comfort.An important note here is that when desired positive affordances and undesired negative affordances are iden-tified early in the design process,then during conceptual design,each concept can be judged relative to all identified positive and negative affordances(see Sect.8).Later, specific affordances must be designed in order to maintain the desired positive affordances while not creating any of the undesired negative affordances.An affordance-based view of the whole design process thus prevents the need to change individual characteristics(such as seat properties in the above example)later when it might be more difficult or expensive to do so.4.4An example of designing a positive AAAAs in AUA,designing an AAA begins with defining and understanding the affordance to be designed.Let us take as our example in this section,the affordance of‘‘turnability’’of gears mentioned earlier.Recall that neither gear need perceive that the other gear(or itself,or the combination of the two)affords turning,since gears,being inanimate objects(i.e.,artifacts),cannot perceive anything.The gears simply afford turning by virtue of their structure.Next,we can consider an extrinsic purpose of this affordance(via the designer),say,of power transmission.The next step will be to identify what physical proper-ties(i.e.,structure)are necessary for both gears such that when put together,the gear pair affords turning,and thereby power transmission.As identified previously,for the affordance of turnability to exist,at a minimum the two gears must have compatible teeth,they must be positioned in the same plane(say,for spur gears,not worm gears)and close enough so that their respective teeth can mesh but not bind,and the two gears must be able to rotate about their respective center axes(Fig.7).Beyond those minimum requirements,the more detailed information the designer has about the intended purpose of the affordance/behavior,the more precise and detailed the list of required physical properties can be.For example,if the designer knows the magnitude of power to be trans-ferred,the loads can be deduced and teeth sizes and profiles。

Chapter8Future TrendsCurrently,the major concerns in the oil industry are in recovering more hydrocar-bons from presently producingfields and improving the success ratio in exploration, development and exploitation.The reservoir team is the new trend in oil company organization and is the key to success in their efforts to enhance their reserves.It has become clear that depth imaging of the subsurface must be done to get better resolution of geometry of the geological structure,fault orientation,as will as structure andfine details of the stratigraphy.A reliable velocity model is required to convert seismic data from the time domain into the depth domain.Such models must represent not only the vertical variation of velocity with depth but also lateral variation caused by heterogeneity of the subsurface rock layers.Using three-component seismic data acquisition both P-wave,SV-wave(vertical component of shear wave)velocities can be determined.This allows various elastic moduli to be determined.The4C systems(four components–the previous three plus hydrophones for OBC measurement),allow many corrections to seismic data to be applied that improve the signal to noise ratio and enhance the seismic resolution.Borehole measurements,by lowering seismic source in a well and receivers in other wells,help to derive velocity tomograms that are used to get other tomograms that represent the spatial variation of petrophysical rock properties such as poros-ity,permeability,clay content,and water saturation to name few.This information will increase our understanding of the reservoir characterization and improve the estimation of the hydrocarbons in place.Advances in3-D seismic data acquisition and computer hardware for data pro-cessing make it more affordable to conduct the3-D seismic surveys,even for ex-ploration efforts to solve the complications of the subsurface geological settings. Lengthy turnaround time for data processing is not an issue,as the giant computing ability and more sophisticated,faster software applications make the acquisition and data processing economically feasible.Tomography,as was discussed in Chap.6,helps in deriving velocity models for pre-stack depth migration and depth conversion.From the elastic moduli,correc-tions for change of seismic velocity with direction of propagation(anisotropy)can be determined and applied in selective areas.The measurements of VSP and long offset VSP can be used to correct the seismic data to the well depth and better image the subsurface geology that will help in locating the next drilling location.M.R.Gadallah,R.Fisher,Exploration Geophysics,227 DOI10.1007/978-3-540-85160-88,c Springer-Verlag Berlin Heidelberg20092288Future Trends 4-D surveys might contribute in increasing the recoverable hydrocarbons as bet-ter understanding of the reservoir performance can be observed through the behavior of the swept front.Pooling leases on the same producing reservoir is not as easy task for the domes-tic oil industry.The merger of oil companies in the past decade will promote the use of3-D time-lapse seismic surveys to monitor secondary and tertiary recoveries projects.Deep offshore-drilling,is another contributor to the increase of reserves.Current crude oil prices,at the time this text was written($50.00–$60.00/bbl),make deep offshore drilling in offshore Angola,South Africa,and the Gulf of Mexico econom-ically feasible The price of crude oil is likely to remain at or above current levels for quite some time.The oil industry is changing very rapidly in all aspect of exploration,develop-ment,and enhanced oil recovery to improve the production efficiency in the current producing oilfields.It is the opinion of the authors that the future of geologists,geo-physicists,land people and engineers is promising and bright.They will face new challenges in an ever-changing industry.With the ups and downs in the industry, many experienced professionals left the industry either for manpower reduction or retirement,resulting in an experience gap among the current employees,intensive training across disciplines is needed to close the gap in communication among the reservoir team members.We are very optimistic that fossil hydrocarbons will be in demand as a source of energy for many decades to come.。

J Huazhong Univ Sci Technol ［Med Sci ］30(3):2010486Expression of Angiotensin Receptors in Aldosterone Ⅱ-producing Adenoma of the Adrenal Gland and Their Clinical Significance *Zhun WU (吴准)1,Dong NI (倪栋)1,Yongji YAN (闫永吉)1,Jun LI (李俊)1,Baojun W ANG (王保军)2,Jinzhi OUYANG (欧阳金枝)3,Guoxi ZHANG (张国玺)4,Xin MA (马鑫)2,Hongzhao LI (李宏召)2,Xu ZHANG (张旭)2#1Department of Urology ,T ongji Hospital,T ongji Medical College,Huazhong University of Science and T echnology,W uhan 430030,China 2Department ofUrology,Chinese PLA General Hospital,Beijing 100086,China 3Department ofEndocrinology,Chinese PLAGeneral Hospital,Beijing 100086,China 4Department ofUrology,First A ffiliated Hospital of Gannan Medical University ,Ganzhou 341000,ChinaHuazhong University of Science and Technology and Springer-V erlag Berlin Heidelberg 2010Summary:The expression of angiotensin Ⅱtype 1receptor (A T 1R)and angiotensin Ⅱtype 2re-ceptor (A T 2R)in aldosterone-producing adenoma (APA)of the adrenal gland was detected,and their relationship with clinical indexes of APA was analyzed.The mRNA expression of A T 1R and A T 2R in 50cases of APA and tissues adjacent to tumors and 12cases of normal adrenal tissues was detected by using reverse transcriptase polymerase chain reaction (RT-PCR).The expression of AT 1R and A T 2R proteins in paraffin-embedded slices of tissue was detected by immunohistochemistry.The ex-pression of A T 1R in adenoma,tissues adjacent to tumor,and normal tissues of the adrenal gland showed no significant differences.The expression of AT 2R in APA tissue was lower than that in nor-mal adrenal gland tissues (P<0.05).Correlation analysis of the mRNA expression level of A T 2R and clinical data from patients demonstrated that AT 2R expression was negatively related to plasma al-dosterone concentration (PAC)(r=－0.467,P<0.05),but positively related with plasma renin activity (PRA)(r=0.604,P<0.05).It is concluded that down-regulation of the AT 2R expression is possibly re-lated with the tumorigenesis of APA.Key words:adrenal gland;aldosterone;adenoma;angiotensin ⅡreceptorPrimary aldosteronism (PA)is a common cause of secondary hypertension,and is associated with a greater number of cardiovascular and renal diseases than essen-tial hypertension [1].The most common clinical subtypes of PA are aldosterone-producing adenoma (APA)and idiopathic hyperaldosteronism (IHA),which accounts for 60%and 30%,respectively [2].As the most common sub-type of P A and usually needing surgical treatment,how-ever,the pathogenesis of AP A still remains unclear .Renin-angiotensin system (RAS)is the principal factor in adjusting the generation of aldosterone.The RAS regula-tion exists not only in the systemic circulation but also in the adrenal cortex based on paracrine [3].Adrenocortical cells can express angiotensin Ⅱtype 1(AT 1R)and type 2receptors (A T 2R).Studies showed that APA demon-strated low reactivity to angiotensin Ⅱ[4].This could be attributed to the changes of the local expression of an-giotensin receptors and the large amount of aldosterone produced through local RAS regulation.Hence,explor-Zhun WU,E-mail:wuzhun2000@#Corresponding author,E-mail:xzhang@ *This project was supported by the National Natural Science F f D Y S f (N 35)ing the existence of abnormal angiotensin Ⅱreceptor expression in APA helps in studying the causes behind PA.This experiment has collected adrenal adenoma and tissues adjacent to tumor from 50patients finally diag-nosed as AP A as well as 12cases of normal adrenal tis-sues.W e then conducted semi-quantitative analysis on A T 1R and A T 2R expression in APA through RT-PCR and immunohistochemistry.1MATERIALS AND METHODS1.1Samples and clinical dataThis study was approved by Ethics Review Com-mittee of Tongji Medical College,Huazhong University of Science and Technology (China),and all patients were given informed consent.We collected adrenocortical adenoma tissues and tissues adjacent to tumor from 50patients who underwent unilateral adrenalectomy in our hospital from January 2007to November 2009.All the patients were finally diagnosed as APA based on endo-crinal examination,computed tomography (CT)of adre-nal glands and (or)adrenal vein sampling (A VS),and histological examination of surgical specimens showed T f 36f f 5y T f 30(4):486-489,2010J Huazhong Uni v Sci Technol ［Med Sci ］DOI 10.1007/s11596-010-0454-0ounds or istinguished oung cholar o China o.072040.an adrenocortical adenoma.his group o patients in-cluded 14m ales and em ales with an average age o 47.ears.he control group had 12sam ples o norm aladrenal tissues from patients receiving radical nephrec-tomy,including 8males and 4females with an average age of 52.4years.All the fresh samples were cut off im-mediately and total RNA was extracted by Trizol and then stored at －80C.Remaining tissues were fixed by formaldehyde,embedded in paraffin,and 4-m thick serial sections were prepared for subsequent immu-nostaining.1.2RT-PCR Semi-Quant itative TestThe agent kit for reverse transcription of total RNA of purified tissue was purchased from TOYOBO (Osaka,Japan).The amplification agent kit TaqMix was pur-chased from Dongsheng Biotech Ltd.(China).The PCR amplimers were synthesized by SANGON Biotech Ltd.(China),and the forward and reverse primers were as follows:5′-TTGCCTGAA TCCTCTTT-3′and 5′-GTGG CTTTGCTTTGTCT-3′for AT 1R gene;5′-GCCTGAA T TTGAAGGAGTGTG-3′and 5′-TAGTAGTGGCAAGG GTGGAGTT-3′for A T 2R gene;5′-AGCGAGCA TCCC CCAAAGTT-3′and 5′-GGGCACGAAGGCTCATCA TT-3′for internal reference β-actin gene.Each 25L assay mix contained of 12.5L 2×PCR TaqMix,2L forward and reverse primers,2ng template DNA.The reaction was subjected to 95°C for 5min,35cycles at 95°C for 30s,annealing temperature (54°C for A T 1R,60°C for A T 2R and β-actin)for 45s,and 72°C for 45s,followed by a final extension at 72°C for 10min.The amplicon size of A T 1R,A T 2R and β-actin was 396,143and 285bp,respectively.A total of 5L products of the amplification process were loaded on 2%agarose gel for electrophoresis and EB dyeing.Afterwards,we examined the results and took photos under ultraviolet in order to measure the band brightness using grayscale analysis software (Image J).1.3Immunohistochemistr yThe SABC method was used to perform the immu-nohistochemical staining.Rabbit anti-human A T 1R,A T 2R polyclonal antibody,and SABC agent kit were purchased from Bios (China).After dewaxing and hydrating the tissue slices,inactive endogenous peroxidase with 3%methyl alcohol-H 2O 2was added,and they were im-mersed into 0.01mol/L citrate buffer solution (pH 6.0).The antigen was restored with microwave for 15min and the slices were rinsed with 0.1mol/L PBS after cooling.After sealing by goat serum,the primary antibody (di-luted to 1:200)was added and the slices were incubated at 4°C overnight.These were then incubated with bio-tin-labeled goat anti-rabbit secondary antibody at room temperature for 30min;these were also incubated at 5%SABC at room temperature for 30min.And then the slices were re-dyed with hematoxylin for 1min and dif-ferentiated with 1%hydrochloric-alcohol for 5s.Finally the slices were dehydrated in xylene solution and sealed with balsam.Upon observing the results under an optical microscope,slices with brown particles in the cell mem-brane were determined as positive immunostaining.The ratio of positive cells in high power fields above 25%was considered as positive protein expression.1.4Stat ist ical AnalysisSPSS13.0statistical analysis software package was used for statistical analysis.The comparisons within y NOV ff yz <5considered statistically significant.2RESULTS2.1RT-P CR Semi-Quantit ative TestThe agarose gel electrophoresis (AGE)results of the RT-PCR test of AT 1R and AT 2R were shown in fig.1and fig.2.The gray ratios AT 1R/β-actin in groups of adenoma,tissues adjacent to tumor,and normal tissues were 0.62±0.12,0.49±0.14and 0.54±0.11,respectively.There were no significant differences in the mRNA ex-pression of AT 1R among all three groups (P>0.05).The gray ratios AT 2R/β-actin in the above 3groups were 0.12±0.01,0.15±0.04and 0.26±0.06,respectively,with the difference being statistically significant between the adenoma and normal adrenal gland groups (P<0.05),but no significant difference was found between group of tissues adjacent to tumor and adenoma group or normal adrenal gland group (P>0.05).The expression of AT 2R mRNA in the adenoma was lower than that in the normal adrenalgland.Fig.1AGE results of AT 1R RT-PCR in normal adrenal tissue (1,2),tissues adjacent to tumor (3,4)and adenoma (5,6).Fig.2AGE results of AT 2R RT-PCR in normal adrenal tissue(1,2),tissues adjacent to tumor (3,4)and adenoma (5,6)2.2Im munohist ochem ist ryIn normal adrenal gland and tissues adjacent to tu-mor,AT 1R and A T 2R proteins were both positively ex-pressed.A T 1R protein immunostaining mainly located in the cell membrane of zona glomerulosa cells,while A T 2R protein immunostaining mainly in the cell membrane of zona reticularis and zona fasciculata cells (figs.3—5).But in the APA tissues,the protein expression of A T 2R was obviously lower than that of A T 1R (figs.6and7).F 3x f T R z (S B ,×;B ×)groups were made b A A.Pearson correlation e i-cient was used to anal e the correlation.A P 0.0wasig.Positive e pression o A 1protein in normal onaglomerulosa A C A:200:400Fig.4Positive expression of AT 2R protein in normal zonafasciculate and zona reticularis (SABC,A:×200;B:×400)Fig.5Positive expression of AT 1R (SABC,A:×400)andAT 2R (SABC,B:×400)protein in tissues adjacent totumorFig.6Positive expression of AT 1R protein in adenoma tissue(SABC,A:×200;B:×400)Fig.7Negative expression of AT 2R protein in adenoma tissue (SABC,A:×200;B:×400)2.3Relations between Expression Level of A T 2R and Clinical Data of PatientsW e also analyzed correlations among relative mRNA expression levels of AT 2R,plasma aldosterone concentration (P AC),plasma renin activity (PRA),plasma potassium concentration (K +),mean arterial blood pressure (MABP)and tumor size.It was found that A T 2R was negatively related to the P AC (r=－0.467,<5),y R (=6,<5)()Table 1Relations between the expression level of AT 2RmRNA and clinical data of patientsRelative expression level of AT 2RClinical indexesr PPAC (ng/dL)－0.470.044PRA (ng/mL h)0.600.006K +(mEq/L)0.220.359MABP (mmHg)－0.090.718Tumor size (cm)－0.320.2143DISCUSSIONThe rennin-angiotensin system (RAS)mediatesseveral classic physiologies including body water and electrolyte homeostasis,blood pressure and generation of aldosterone.These functions appear to be mediated by A T 1R and A T 2R subtype system.In the adrenal cortex,renin,angiotensinogen,AT 1R and A T 2R together com-pose a local angiotensin Ⅱsystem regulated independ-ently from the circulating RAS,which is suggested to have a local physiological function [5].The major subtype of angiotensin receptor for adults is A T 1R,which is the classic G protein-coupled transmembrane receptor.Through the signal pathways of A T 1R,the angiotensin Ⅱcan complete functions such as contracting vascular smooth muscles,secreting aldosterone,and enhancing the hypertrophy and proliferation of cells [6,7].At present,rare studies on the expression of angiotensin receptors in AP A have been conducted.International reports on A T 1R expression in adrenocortical adenoma reveal different viewpoints.The expression of A T 1R is normal according to Opocher [8]and Marek [9],but Nawata [10]and Barbara [11]found A T 1R expression was much higher in adenoma than in normal adrenal gland.In our study,we compared the expression of A T 1R within adenoma and tissues ad-jacent to the tumor of 50AP A patients and 12normal adrenal tissues.Both RT-PCR and immunohistochemis-try demonstrated that the expression variance of A T 1R in the three sample groups showed no statistical signifi-cance.The A T 2R protein is also belong to seven trans-membrane G protein-coupled super-family receptors and has 34%homology compared with AT 1R [12].Grady [13]found that A T 2R is mainly distributed in embryonic tis-sues.After birth,A T 2R is reduced or disappears rapidly in most tissues and organs.In adults,A T 2R expression is restricted to only a few organs,such as brain tissues,adrenal glands,ovary,and heart [13].It has been shown that stimulation of A T 2R induces a decrease in intracel-lular cGMP levels in neuronal cells [14].However,the physiological roles of AT 2R in adrenal gland remain to be clarified.Current studies have shown that AT 1R and A T 2R possess counter-regulatory roles in many aspects.For example,AT 1R makes the blood vessel contract,while A T 2R makes the blood vessel dilate [15].In addition,A T 1R can help promote cell proliferation,while AT 2R can hinder the same process [16].A T 2R appear to counter-balance the effects of angiotensin Ⅱmediated by A T 1R,including cell growth,vasoconstriction,cardiovascular fibrosis and so on [16,17].Interestingly,A T 2R seems to be f F y,T R X ,f f f y P 0.0but positivel related to the P A r 0.04P 0.0table 1.connected with the incidence o AP A.irstl A 2gene is located on the chromosome which complements the act that the incidence o APA is in luenced b genderdifference (the amount of female APA patients is about three times more than male patients)[2].Moreover,Wolf et al [18]have found that angiotensin inhibits cell growth through A T 2R.They found the quantity of A T 2R was gradually reduced with the adrenal gland growing since embryonic period,which signifies that the reduced A T 2R expression level might promote the growth and differen-tiation of adrenocortical cells.Horiuchi et al [19]found that angiotensin Ⅱreceptors promote cell apoptosis in the PC12W and R3T3cell lines.In addition,the A T 2receptor antagonist can inhibit cell apoptosis but A T 1receptor antagonist (ARB)is invalid,which indicates that A T 2R can mediate cell apoptosis.In this study,we compared the expression of AT 2R in three groups of samples:adenoma,tissues adjacent to tumor,and normal adrenal gland.Both RT-PCR and immunohistochemical staining showed that AT 2R ex-pression in APA was significantly lower than that in normal adrenal gland.We also analyzed correlations within relative mRNA expression levels of AT 2R and clinical data of patients,in which we found that AT 2R was negatively related to the PAC but positively related to the PRA.We can assume that the reduced expression level of AT 2R in adrenocortical cell is potentially related with the tumorigenesis and function of APA.Down-regulation of AT 2R leading to deficient counterbalance to the effects of angiotensin Ⅱmediated by AT 1R finally could promote adrenal zona glomerulosa cell prolifera-tion and aldosterone hypersecretion.However,these possible mechanisms need further study on cellular or animal levels.REFERENCES 1Milliez P,Girerd X,Plouin PF,et al.Evidence for an in-creased rate of cardiovascular events in patients with pri-mary aldosteronism.J Am Coll Cardiol,2005,45(8):1243-12482Stewart PM.Mineralocorticoid ncet,1999,353(9161):1341-13473Fallo F,Pistorello M,Pedini F,et al.In vitro evidence forlocal generation of renin and angiotensin Ⅱ/Ⅲimmuno-reactivity by the human adrenal gland.Acta Endocrinol (Copenh),1991,125(3):319-3304Brown RD,Wisgerhof M,Carpenter PC,et al.Adrenalsensitivity to angiotensin Ⅱand undiscovered aldosterone stimulating factors in hypertension.J Steroid Biochem,1979,11(1C):1043-10505Ehrhart-Bornstein M,Hinson JP,Bornstein SR,et al.In-traadrenal interactions in the regulation of adrenocorticalsteroidogenesis.Endocr Rev,1998,19(2):101-1436de Gasparo M,Catt KJ,Inagami T,et al.Internationalunion of pharmacology.XXIII.The angiotensin Ⅱrecep-tors.Pharmacol Rev,2000,52:415-4727McEwan PE,Vinson GP,Kenyon CJ.Control of adrenal cell proliferation by A T1receptors in response to angio-tensin Ⅱand low-sodium diet.Am J Physiol,1999,276(2Pt 1):E303-3098Opocher G ,Rocco S,Cimolato M,et al.Angiotensin Ⅱreceptors in cortical and medullary adrenal tumors.J Clin Endocrinol Metab,1997,82(3):865-8699Pawlikowski M,Winczyk K,Sledz B.Immunohisto-chemical detection of angiotensin receptors A T1and AT2in adrenal tumors.Folia Histochem Cytobiol,2008,46(1):51-5510Nawata H,Takayanagi R,Ohnaka K,et al.Type 1angio-tensin Ⅱreceptors of adrenal tumors.Steroids,1995,60(1):28-3411Schubert B,Fassnacht M,Beuschlein F,et al.Angiotensin Ⅱtype 1receptor and ACTH receptor expression in hu-man adrenocortical neoplasms.Clin Endocrinol (Oxf),2001,54(5):627-63212Tissir F,Riviere M,Guo DF,et al.Localization of the genes encoding the three rat angiotensin Ⅱreceptors,Agtr1a,Agtr1b,Agtr2,and the human AGTR2receptor respectively to rat chromosomes 17q12,2q24and Xq34,and the human Xq22.Cytogenet Cell Genet,1995,71(1):77-8013Grady EF,Sechi LA,Griffin CA,et al.Expression of AT2receptors in the developing rat fetus.J Clin Invest,1991,88(3):921-93314Israel A,Stromberg C,Tsutsumi K,et al.Angiotensin Ⅱreceptor subtypes and phosphoinositide hydrolysis in rat adrenal medulla.Brain Res Bull,1995,38(5):441-44615Carey RM,Wang ZQ,Siragy HM.Role of the angiotensin type 2receptor in the regulation of blood pressure and re-nal function.Hypertension,2000,35(1Pt 2):155-16316Nakajima M,Hutchinson HG,Fujinaga M,et al.The an-giotensin Ⅱtype 2(AT2)receptor antagonizes the growth effects of the AT1receptor:gain-of-function study using gene transfer.Proc Natl Acad Sci U S A,1995,92(23):10663-1066717Inagami T,Eguchi S,Numaguchi K,et al.Cross-talk be-tween angiotensin Ⅱreceptors and the tyrosine kinases and phosphatases.J Am Soc Nephrol,1999,10(Suppl 11):S57-S6118Wolf G,Harendza S,Schroeder R,et al.Angiotensin Ⅱ's antiproliferative effects mediated through AT2-receptorsdepend on down-regulation of b Invest,2002,82(10):1305-131719Horiuchi M,Hayashida W,Kambe T,et al.Angiotensintype 2receptor dephosphorylates Bcl-2by activating mi-togen-activated protein kinase phosphatase-1and induces apoptosis.J Biol Chem,1997,272(30):19022-19026(Received March 17,2010)。

ORIGINAL PAPERA mutation in a coproporphyrinogen III oxidase gene confers growth inhibition,enhanced powdery mildew resistance and powdery mildew-induced cell death in ArabidopsisChuan-yu Guo •Guang-heng Wu •Jin Xing •Wen-qi Li •Ding-zhong Tang •Bai-ming CuiReceived:27November 2012/Revised:7February 2013/Accepted:20February 2013ÓSpringer-Verlag Berlin Heidelberg 2013AbstractKey message A gene encoding a coproporphyrinogen III oxidase mediates disease resistance in plants by the salicylic acid pathway.Abstract A number of genes that regulate powdery mil-dew resistance have been identified in Arabidopsis,such as ENHANCED DISEASE RESISTANCE 1to 3(EDR1to 3).To further study the molecular interactions between the powdery mildew pathogen and Arabidopsis,we isolated and characterized a mutant that exhibited enhancedresistance to powdery mildew.The mutant also showed dramatic powdery mildew-induced cell death as well as growth defects and early senescence in the absence of pathogens.We identified the affected gene by map-based cloning and found that the gene encodes a coproporphy-rinogen III oxidase,a key enzyme in the tetrapyrrole biosynthesis pathway,previously known as LESION INI-TIATION 2(LIN2).Therefore,we designated the mutant lin2-2.Further studies revealed that the lin2-2mutant also displayed enhanced resistance to Hyaloperonospora ara-bidopsidis (H.a.)Noco2.Genetic analysis showed that the lin2-2-mediated disease resistance and spontaneous cell death were dependent on PHYTOALEXIN DEFICIENT 4(PAD4),SALICYLIC ACID INDUCTION -DEFICIENT 2(SID2),and NONEXPRESSOR OF PATHOGENESIS -RELATED GENES 1(NPR1),which are all involved in salicylic acid signaling.Furthermore,the relative expres-sion levels of defense-related genes were induced after powdery mildew infection in the lin2-2mutant.These data indicated that LIN2plays an important role in cell death control and defense responses in plants.Keywords Arabidopsis ÁLIN2ÁCell death ÁDisease resistance ÁSA pathwayAbbreviations LIN2LESION INITIATION 2CPO Coproporphyrinogen III oxidaseEDR1ENHANCED DISEASE RESISTANCE 1PAD4PHYTOALEXIN DEFICIENT 4SID2SALICYLIC ACID INDUCTION-DEFICIENT 2NPR1NONEXPRESSOR OF PATHOGENESIS-RELATED GENES 1EIN2ETHYLENE INSENSITIVE 2COI1CORONATINE-INSENSITIVE 1Communicated by Y.Lu.C.Guo and G.Wu contributed equally to this work.Electronic supplementary material The online version of this article (doi:10.1007/s00299-013-1403-8)contains supplementary material,which is available to authorized users.C.Guo ÁJ.Xing ÁW.Li ÁB.Cui (&)College of Life Science,Shihezi University,Shihezi 832003,Xinjiang,People’s Republic of China e-mail:baimingc@ C.Guoe-mail:siweiguosongyao@ J.Xinge-mail:xingjin_1021@ W.Lie-mail:susu109li@G.Wu ÁD.Tang (&)State Key Laboratory of Plant Cell and ChromosomeEngineering,Institute of Genetics and Developmental Biology,Chinese Academy of Sciences,Beijing 100101,People’s Republic of China e-mail:dztang@ G.Wue-mail:wuguangheng@Plant Cell RepDOI 10.1007/s00299-013-1403-8KEG KEEP ON GOINGSR1SIGNAL RESPONSIVE1LMMs Lesion mimic mutantsPCD Programmed cell deathSA Salicylic acidJA Jasmonic acidET EthyleneHR Hypersensitive responseSAR Systemic acquired resistanceDAB DiaminobenzidineIntroductionPlants are continually exposed to a variety of potential pathogens,and to survive they have evolved sophisticated defense responses to the majority of these pathogens.The responses include cell wall strengthening,the hypersensi-tive response(HR),increased expression of defense-related genes,the oxidative burst and production and delivery of antimicrobial compounds to the site of infection(Green-berg1997;Mehdy1994;Staskawicz et al.1995).The HR is characterized by rapid death of plant cells directly in contact with,or close to,the pathogen,which leads to an increase in the biosynthesis of SA,the expression of PATHOGENESIS-RELATED(PR)genes and the induction of systemic acquired resistance(SAR)(Durrant and Dong 2004;Glazebrook2005;Ryals et al.1996;Ward et al. 1991;Yalpani et al.1991).Plant disease resistance and HR-induced cell death are mediated by several signaling molecules,including SA, jasmonic acid(JA),and ethylene(ET)(Dong1998; Glazebrook2001;Kunkel and Brooks2002;Loake and Grant2007;Thomma et al.2001).Generally,the SA pathway is critical for resistance to biotrophic pathogens, while the JA/ET pathway functions in resistance to nec-rotrophic pathogens.By genetic screening,a number of crucial components that function in the SA pathway have been discovered,including SID2,PAD4,and NPR1.The loss of function of these genes leads to enhanced suscep-tibility to various pathogens(Cao et al.1997;Delaney et al. 1995;Jirage et al.1999;Nawrath et al.2002;Parker et al. 1996;Wildermuth et al.2001).Similarly,plants carrying the NahG transgene,which degrades endogenous SA,dis-play a high susceptibility to a variety of pathogens(Gaff-ney et al.1993).In addition,JA/ET signal pathway related genes,such as ETHYLENE INSENSITIVE2(EIN2)and CORONATINE-INSENSITIVE1(COI1),play important roles in the defense responses to necrotrophic pathogens (Alonso et al.1999;Feys et al.1994).Over the last decade,substantial efforts have been made to identify genes involved in the regulation of disease resistance and HR-induced cell death.A number of Ara-bidopsis mutants showing enhanced disease resistance have been identified.For example,edr1(enhanced disease resistance1),edr2and edr3mutants exhibit enhanced disease resistance to powdery mildew and show mildew-induced cell death.Defense-related genes are not constitu-tively expressed in edr1,edr2and edr3mutants,but are more strongly and rapidly induced when these mutants are infected with powdery mildew(Frye et al.2001;Nie et al. 2011;Tang et al.2005a,b,2006;Vorwerk et al.2007). Double-mutant analysis revealed that mutations that blocked SA-associated defense responses could suppress edr1-mediated resistance.In contrast,mutations in EIN2 and COI1did not suppress edr1-mediated resistance,indi-cating ET/JA-induced responses were not required for edr1 resistance.In addition,the ein2mutation suppressed the edr1-mediated ethylene-induced senescence phenotype,but mutations in SID2,PAD4,EDS1,or NPR1did not,sug-gesting that EDR1functions at a point of crosstalk between ethylene and salicylic acid signaling that impinges on senescence and cell death(Tang et al.2005a,b).Tran-scriptome analyses on wild-type and edr1plants inoculated with powdery mildew(Golovinomyces cichoracearum) revealed that endomembrane system,defense and reactive oxygen species(ROS)production-related genes were induced early and maintained for at least4days after infection in the edr1mutant,and that expression level of genes related with protein kinases was also improved early in edr1mutant(Christiansen et al.2011).Furthermore,by suppressor screening,a number of genes involved in the edr1and/or edr2resistance pathway have been identified.A missense KEEP ON GOING(KEG,encodes an E3ubiquitin ligase)mutation fully suppressed all edr1-associated phe-notypes(Wawrzynska et al.2008),suggesting that protein degradation mediated by KEG plays a role in EDR1sig-naling.Recently,it was shown that KEG may recruit EDR1 to the trans-Golgi network/early endosome vesicles,indi-cating that EDR1may play a role in vesicle trafficking(Gu and Innes2011).A mutation in Arabidopsis HPR1,a component of the THO/TREX complex,suppresses edr1-mediated resistance to the powdery mildew pathogen and mildew-induced cell death,but enhanced the ethylene-induced senescence phenotype of edr1(Pan et al.2012).A gain-of-function mutation in SIGNAL RESPONSIVE1 (SR1),which encodes a calmodulin-binding transcription activator,suppresses all edr2phenotypes,including pow-dery mildew resistance,mildew-induced cell death,and ethylene-induced senescence.Also,enhanced ethylene-induced senescence in sr1-1is suppressed by ein3(Nie et al. 2012).A mutation in the gene encoding RPN1a,a subunit of the26S proteasome,suppresses edr1-,edr2-and pmr4-associated disease resistance phenotypes.Accumulation of RPN1a is affected by SA,and the SA accumulation in anPlant Cell Reprpn1a mutant is lower than in wild type upon Pto DC3000 infection(Yao et al.2012).In addition to the edr mutants,there is another group of disease resistance mutants with spontaneous cell death, which are known as lesion mimic mutants(LMMs).They exhibit either constitutive or unregulated cell death forma-tions that resemble HR after pathogen infection.Lesion mimic is considered the result of programmed cell death (PCD)(Dangl et al.1996).Based on different mechanisms involved in controlling cell death,LMMs could be defined into two classes:initiation mutants and propagation mutants (Lorrain et al.2003).Initiation mutants form localized necrotic spots of determinate size,such as mutants cpn1 (copine),lin2(lesion initiation),hrl1(hypersensitive response-like lesions)and hlm1(HR-like lesion mimic), which display relatively low levels of cell death compared to propagation mutants(Balague et al.2003;Devadas et al. 2002;Ishikawa et al.2001;Jambunathan et al.2001). Propagation mutants have defects that do not allow them to control cell death once it has started,such as the acd (accelerated cell death)mutants and the lsd1(lesion sim-ulating disease resistance response1)mutant,which develop chlorotic or tissue-specific lesions(Dietrich et al. 1994;Greenberg and Ausubel1993;Greenberg et al.1994). Since thefirst LMM was isolated from maize(Hoisington et al.1982),many mutants exhibiting spontaneous cell death were identified in Arabidopsis(Dietrich et al.1994; Greenberg and Ausubel1993),barley(Wolter et al.1993), rice(Takahashi et al.1999),and tomato(Tang et al.1999).One useful approach to the investigation of mechanisms involved in regulating disease resistance and cell death in plants is the isolation of edr1-like mutants or LMMs that display a misregulation of cell death formation induced either by pathogens or spontaneously(Mur et al.2008).Here,we characterized a mutant that shows enhanced resistance to powdery mildew and exhibits mildew-induced cell death.The mutant also showed enhanced resistance to oomycete pathogen H.a.Noco2.We identified the corre-sponding gene by map-based cloning and found that the mutant phenotypes were caused by a mutation in LIN2, which encodes a coproporphyrinogen III oxidase.There-fore,we named the mutant lin2-2.We showed that LIN2 acted as a negative regulator of plant defense responses and cell death,and the lin2-2-mediated defense and spontane-ous cell death is SA dependent but JA/ET independent. Materials and methodsPlant materials and growth conditionsArabidopsis seeds were surface sterilized by10%bleach and chilled at4°C for2days,then sowed on plates with half-strength Murashige and Skoog(1/2MS)medium plus 0.8%agar and placed in a growth room at21°C with a 9-h-day/15-h-night photoperiod(short photoperiod)(long photoperiod conditions for identification of photoperiod-dependent spontaneous cell death phenotype).Seedlings were transplanted into a mixture of commercial potting soil and perlite(3:1)in a growth chamber with a9-h-day/15-h-night photoperiod at21°C and80%relative humidity.For seed set,the plants were moved to a growth room at21°C with a16-h-day/8-h-night photoperiod(long photoperiod).Mutant screeningArabidopsis Columbia-0(Col-0)seeds were mutagenized with0.2%ethyl methylsulfonate(EMS).The M2genera-tions were grown in a growth room at21°C with a9-h-day/ 15-h-night photoperiod.Four-week-old plants were inocu-lated with G.cichoracearum UCSC1and scored for resis-tance at8days post-infection(dpi)(Nie et al.2011).Plants displaying powdery mildew resistance with no visible fungal spores and pathogen induced or spontaneous cell death phenotypes,were selected.Approximately55,000M2 plants derived from7,500M1plants were inoculated. Mapping and cloningThe lin2-2mutant with powdery mildew resistance and spontaneous cell death phenotype was selected,and the F2 plants derived from a cross of the lin2-2mutant(Col-0 background)9Ler(Landsberg erecta)were used for map-ping.Then,the powdery mildew resistance and spontaneous cell death phenotype plants were selected for rough mapping with simple sequence length polymorphism(SSLP)markers. The mutant was localized between molecular markers T7I23 and F10B6on chromosome I.To narrow down the region,F3 plants derived from homozygous F2plants that were segre-gated into powdery mildew resistance and spontaneous cell death plants were used.We developed our own markers at this interval using Monsanto Col-0and Landsberg polymorphism data.Five-hundred F3plants were scored and the mutation was mapped to an85kb interval.Candidate genes in this region were sequenced to identify the mutation.Gene annotation and sequences analysisThe85kb region was analyzed in the Arabidopsis genome automated annotation database(/). All of the possible target ORFs were amplified using KOD polymerase(Toyobo Company,Japan)and purified using DNA gel extraction kit.Then,all the PCR fragments were sequenced by Beijing Genomics Institute of China. Sequencing results were aligned using Vector NTI Suite10 (/).Plant Cell RepPathogens infectionPowdery mildew(Golovinomyces cichoracearum UCSC1) strain was maintained on highly susceptible pad4-1mutant plants(Tang and Innes2002).Four-week-old plants were inoculated by transferring conidia(asexual spores)from pad4-1plants as described previously(Wang et al.2011). The disease or resistance phenotypes were scored8days after infection by assessment of spore production or lesion formation(Nie et al.2011).After seedlings were transplanted into a mixture of commercial potting soil and perlite(3:1)in a growth chamber with a16-h-day/8-h-night photoperiod(long photoperiod)at21°C and80%relative humidity for 2weeks,infection of H.a.Noco2was performed on these seedlings by spraying with a H.a.Noco2spore suspension at a concentration of59104spores per milliliter of water. Then the plants were kept at18°C in12h light/12h dark cycles with95%humidity,and the infection was scored 7days after inoculation by counting the number of coni-diospores per gram of leaf tissue using a hemocytometer as previously described(Li et al.2010).Staining and microscopyFungal structures and dead cells were stained using trypan blue(Koch and Slusarenko1990).Diaminobenzidine (DAB)staining was performed according to the method of Hans Thordal-Christensen et al.(1997)to detect the accumulation of ROS.Leaves stained with DAB were examined with a light microscope.Ethylene-induced senescence assayFive-week-old plants were treated with100l L/L ethylene in a sealed chamber for3days.Chlorophyll was extracted and measured following Frye et al.(2001).Complementation of the lin2-2mutationThe following primers50-GGGAATTCCCACCTGAGTC TGATGAGCCTGATGGTT-30and50-GGGAATTCCCG GATGACATGGCAGAACTTCAC-30were used for PCR to isolate the full-length genomic sequence,which included 1,745bp upstream of the start codon and428bp downstream of the stop codon of AT1G03475.1.These primers contained the restriction enzyme sites EcoR I at their50end for sub-sequent cloning.The PCR product wasfirst cloned into pMD18-T vector and then introduced into the plasmid pCAMBIA1300.The LIN2genomic clone construct was then introduced into Agrobacterium tumefaciens strain GV3101, and the lin2-2mutant plants were transformed using thefloral dip method(Clough and Bent1998).Transgenic plants were selected by growth on1/2MS containing0.8%agar and50mg/L hygromycin.Trans-formants were transplanted into soil7days after germi-nation and phenotype was observed.Construction of double mutantsThe following mutants or transgenic plants were crossed with lin2-2to construct double mutants:edr1(Frye and Innes1998),edr2(Tang et al.2005a,b),pad4-1(Glaze-brook et al.1997),sid2-2(Wildermuth et al.2001),npr1/ nim1(in Ws background)(Delaney et al.1995),coi1-1 (Xie et al.1998),ein2-1(Guzman and Ecker1990),and the transgenic line NahG(in Ler background)(Bowling et al. 1994).Quantitative real-time RT-PCRPlants were grown and infected with G.cichoracearum as described above.Leaves were detached from plants at different time points.Total RNA was isolated using the RNeasy Kit,followed by reverse transcription to cDNA using SuperScript reverse transcriptase(Invitrogen).The primers used in quantitative real-time RT-PCR are listed in Supplementary Table1.Construction of the LIN2promoter::GUS reporterand GUS activity assayA1.88-kb50-fragment of LIN2was amplified using primers 50-GGGGACAAGTTTGTACAAAAAAGCAGCTTCTCT GGTTTGTGTCGTA-30and50-GGGGACCACTTTGTA CAAGAAAGCTGGGTCCAAGGCGAGATTAGGTTT-30 from genomic DNA of a wild-type plant.The promoter sequence wasfirst cloned into pDONR207vector and then introduced into the plasmid pMDC163to fuse with the GUS reporter gene,and the construct was transformed into Col-0plants.A GUS activity assay was performed as previously described(Tang et al.2005a,b).ResultsThe lin2-2mutant displays photoperiod-dependent spon-taneous cell death,enhanced powdery mildew resistance and powdery mildew-induced cell death.To identify mutants that display enhanced disease resistance to powdery mildew,we mutagenized Col-0 seeds with0.2%ethyl methanesulfonate(EMS).The M2 plants were inoculated with Arabidopsis powdery mildew Golovinomyces cichoracearum(G.cichoracearum)strain UCSC1and scored for disease resistance phenotypes at 8days post-infection(dpi).Plants displaying enhancedPlant Cell Reppowdery mildew resistance were selected.Here,we describe one mutant with enhanced powdery mildew resistance and powdery mildew-induced cell death,desig-nated lin2-2based on the subsequent characterization.To characterize the growth phenotype of the lin2-2 mutant,we grew wild-type and lin2-2plants on plates under long photoperiod or short photoperiod conditions. One week later,the lin2-2mutant under long photoperiod (long day,LD)conditions displayed obvious growth defects with shorter roots and smaller leaves than that of Col-0,and it showed spontaneous lesions on the leaves (Fig.1a,upper lanes),while no differences were observed between lin2-2and Col-0under the short photoperiod (short day,SD)conditions(Fig.1a,down lanes).Seedlings grown under the short photoperiod7days after germina-tion were transplanted into soil.After2weeks,the lin2-2 mutant displayed growth defects and was significantly different from that of Col-0,with obvious spontaneous cell death,growth inhibition and pale green leaves(Fig.1b and Supplemental Fig.S1b).To further assess the lin2-2growth phenotype,we grew the plants over6weeks under the long photoperiod conditions,and the lin2-2was significantly weaker with more severe early senescence,fewer tillers and shorter siliques than that of Col-0(Supplemental Fig. S1a–d).To examine powdery mildew resistance,4-week-old plants grown under the short photoperiod conditions were inoculated with the powdery mildew pathogen G.cichor-acearum UCSC1.While a large number of conidiophores (asexual reproductive structures)were produced on the leaves of Col-0at7dpi,lin2-2exhibited a resistance phenotype with no visible powdery mildew growth,but with dramatic lesion development at the infection sites (Fig.1c).Trypan blue staining showed that there were a large number of fungal hyphae and conidiophores but no visible lesions were observed on the Col-0leaves.In contrast,the number of fungal hyphae and conidiophores was significantly fewer,with more and larger patches of dead mesophyll cells on the lin2-2leaves(Fig.1d).The conidiophore numbers in wild-type and lin2-2plantsat Fig.1The lin2-2mutant displays photoperiod-dependent spontane-ous cell death and enhanced disease resistance.a Phenotype ofseedlings7days after germination under long-day(LD)conditions(upper lane)and short-day(SD)conditions(down lane),respectively.Bar3mm.b Phenotype of seedlings2weeks after transplanted insoil under short-day conditions.Bar3mm.c Single representativeleaves removed from plants and photographed at7dpi.Bar3mm.d Leaves from plants at7dpi stained with trypan blue.Bar100l m.e Quantification of fungal growth in plants at7dpi by calculating thenumber of conidiophores per colony.Results originate from oneexperiment and bars represent mean and standard deviation from25colonies.Statistically significant differences are indicated by asterisk(P\0.05;Student’s t test).The experiment was repeated three timeswith similar results.f DAB staining for hydrogen peroxide at2dpi.Bar100l mPlant Cell Rep7dpi was consistent with the observations in the trypan blue staining (Fig.1e).Accumulation of hydrogen peroxide (H 2O 2)often occurred around cells that undergo cell death (Fellbrich et al.2002).To assess H 2O 2accumulation in the wild-type and the lin2-2mutant,we examined H 2O 2accumulation at 2days after infection with G.cichoracearum .As shown in Fig.1f,the H 2O 2level in the lin2-2mutant was much higher than that in Col-0.The lin2-2mutant shows enhanced resistance to H.a.Noco2Previously,it was shown that PR1was constitutively expressed in lin2(Ishikawa et al.2001)and that lin2-2exhibited enhanced resistance to powdery mildew.Thus,we speculate that lin2-2might exhibit resistance to a broad spectrum of pathogens.We inoculated lin2-2with the virulent oomycete pathogen H.a.Noco2.As shown in Fig.2,the wild type was susceptible to H.a.Noco2,while lin2-2was more resistant (Fig.2).The lin2-2mutant exhibits enhanced ethylene-induced senescenceA previous report revealed that several powdery mildew resistant mutants exhibited enhanced ethylene-induced senescence (Frye et al.2001;Tang et al.2005a ,b ;Wang et al.2011).To examine whether senescence of lin2-2could be accelerated by ethylene,wild type and lin2-2were exposed to 100l L/L ethylene for pared to wild type,lin2-2exhibits enhanced ethylene-induced senescence (Fig.3a).Additionally,the chlorophyll level was significantly lower in lin2-2than in the wild type when treated with ethylene for 3days (Fig.3b).Identification of the lin2-2mutationTo identify the gene for the lin2-2mutant,map-based cloning was employed using an F 2population obtained from a cross of lin2-2(Col-0background)and Landsberg erecta (Ler).Forty-six lesion mimic plants were selected in the F 2population for rough mapping,and the mutation was mapped to a region close to SSLP marker F12K11on chromosome I.To further narrow the mapping region,we used a large number of F 3plants derived from F 2plants displaying powdery mildew resistance and the spontaneous cell death phenotype.Fine mapping localized the corre-sponding gene to an 85kb region (Fig.4a),which con-tained 40genes.In this region,we found a single point mutation C653T in AT1G03475.1,which caused a mis-sense mutation from Thr to Ile in the lin2-2mutant (Fig.4b).To confirm the mutant phenotypes we observed in lin2-2were caused by this mutation,we made a full-length LIN2genomic clone under the control of its native promoter,and transformed it into the lin2-2mutant.The genomic DNA of LIN2complemented the lin2-2pheno-types of spontaneous cell death,early senescence,growth inhibition and pale green leaves (Fig.4c).The LIN2genomic clone also complemented the lin2-2-mediated powdery mildew resistance phenotype,as no obvious cell death and a large number of fungal hyphae and conidio-phores were observed in the transgenic plant leaves (Fig.4d).The LIN2gene encodes coproporphyrinogen III oxi-dases (CPO),which exist widely in living organisms,including both lower forms of life,such as the unicellular microalgae Botryotinia fuckeliana and the small arthropod Camponotus floridanus ,and higher living organisms,such as Homo sapiens ,suggesting that CPO has an important function in different organisms (Supplemental Fig.S2a).Furthermore,the alignment of LIN2with its homologs showed that the CPO proteins were highly conserved in different species,especially in the C-terminal region (Supplemental Fig.S2b).A protein sequence alignment analysis revealed that the missense mutation site was highly conserved from unicellular living organisms to human (Supplemental Fig.S2b).The spontaneous cell death of lin2-2is dependent on SA,but not JA/ETTo investigate the role of SA,JA and ET in spontaneous cell death and the growth defect of lin2-2,lin2-2was crossed with several mutants that have defects in the SA or JA/ET pathway.The lin2-2pad4and lin2-2npr1showed wild-type-like growth and displayed no spontaneous cell death,indicating that pad4and npr1suppressed lin2-2-mediated growth defects and spontaneous celldeathFig.2lin2-2Plants displayed enhanced resistance to H.a.Noco2.Two-week-old seedlings growing in a growth room with long photoperiod were sprayed with H.a.Noco2(59104spores/mL)and kept at 18°C in 12h light/12h dark cycles with 95%humidity.Infection was scored at 7days post-inoculation by counting the number of conidia spores.The bars represent mean and standard deviation from 25plants.The asterisk indicates a significant difference from wild type (P \0.05;Student’s t test).The experiment was repeated three times with similar resultsPlant Cell Repphenotypes.However,the sid2-2mutation could inhibit spontaneous cell death,but not restore growth defects in lin2-2.Similarly,the NahG transgene inhibited necrotic lesions but partially restored growth defects in lin2-2.In contrast,ein2-2and coi1-1did not suppress any phenotype of lin2-2(Fig.5).These data indicated that the phenotypes of lin2-2are partially dependent on SA,but not JA/ET.Powdery mildew resistance in lin2-2requires the SA pathwayPlant defense responses are regulated through a complex network of signaling pathways that involve three signaling molecules:SA,JA and ET (Kunkel and Brooks 2002).To gain more insight into how LIN2regulates defense responses in Arabidopsis,we assessed the roles of SA,JA and ET in lin2-2-mediated disease resistance by double-mutant analysis.Four-week-old single and double mutants were inoculated with the powdery mildew pathogen.As shown in Fig.6a,pad4,sid2,coi1and ein2could notinhibit powdery mildew-induced cell death of lin2-2.However,trypan blue staining revealed that lin2-2pad4and lin2-2sid2are less resistant than lin2-2,indicating that lin2-2-mediated powdery mildew resistance is SA signal-ing dependent (Fig.6b).To quantify the disease resistance of each genotype,the number of conidiophores in infected leaves at 7dpi was calculated.The number of conidio-phores in the lin2-2mutant was strikingly lower than the wild type,while pad4and sid2supported more conidio-phore growth than the wild type.The levels of mildew growth in the coi1and ein2mutants were similar to that of wild type (Fig.6c).In the double mutants,lin2-2pad4and lin2-2sid2,the number of conidiophores was higher than in Col-0(Fig.6c).In contrast,the lin2-2coi1and lin2-2ein2double mutants supported similar numbers of conidio-phores as lin2-2(Fig.6c).Taken together,the results indicated that lin2-2-mediated powdery mildew resistance is SA signaling pathway dependent and JA/ET signaling pathway independent,while the powdery mildew-induced cell death is SA and JA/ET pathway independent.The edr1mutation enhances the lin2-2phenotype Previous studies showed that edr1exhibited enhanced resistance to powdery mildew and pathogen-induced cell death (Frye and Innes 1998;Tang et al.2005a ,b ).To further assess lin2-2-mediated resistance to powdery mil-dew,we crossed lin2-2with edr1and examined the lin2-2edr1phenotypes.In the absence of the pathogen,more obvious cell death occurred,as did dramatic growth inhi-bition and early senescence,in 6-week-old lin2-2edr1plants than in the lin2-2single mutant (Fig.5).To inves-tigate whether edr1affects powdery mildew resistance in lin2-2,conidiophore formation was measured in lin2-2and lin2-2edr1.The lin2-2edr1double mutants supported significantly fewer conidiophores than lin2-2,indicating that edr1enhanced lin2-2-mediated powdery mildew resistance (Fig.6b,c).Expression of PR genes in lin2-2To analyze the expression of defense-related genes in the lin2-2mutant,quantitative real-time PCR was performed to monitor relative expression of three PR genes,PR1,PR2and PR5,at various time points after inoculation with G.cichoracearum .As shown in Fig.7,PR1,PR2and PR5transcript levels were relatively low in Col-0,while the transcriptional accumulation of these genes,especially PR1,in lin2-2was high before inoculation.The tran-scriptional accumulation of PR genes rapidly increased after inoculation.PR1reached a peak at 5dpi,while the transcription levels of PR2and PR5peaked at 3dpi in lin2-2plants (Fig.7).Fig.3lin2-2Plants exhibited enhanced ethylene-induced senes-cence.Five-week-old plants were treated with 100l L/L ethylene for 3days in a sealed chamber in a growth room under short photoperiod conditions.a Plants were treated with 100l L/L ethylene for 3days.Note that,on the wild-type plants,only old leaves show senescence,whereas in lin2-2plants,senescence occurred in much younger leaves.Bar 5mm.b Quantification of chlorophyll content in wild-type and lin2-2leaves.The leaves were detached (excluding the two oldest true leaves),and chlorophyll was extracted and quantified as previously described.The bars represent the mean and standard deviation from 15plants.The asterisk indicates a significant difference from wild-type (P \0.05;Student’s t test).The experi-ments were repeated three times with similar resultsPlant Cell Rep。

Appl Phys A(2011)105:445–451DOI10.1007/s00339-011-6577-8Effect of nanostructures on evaporation and explosive boiling of thin liquidfilms:a molecular dynamics studyA.K.M.M.Morshed·Taitan C.Paul·Jamil A.KhanReceived:2May2011/Accepted:17August2011/Published online:14September2011©Springer-Verlag2011Abstract Molecular dynamics simulations have been em-ployed to investigate the boiling phenomena of thin liq-uidfilms adsorbed on a nanostructured solid surface.The molecular system was comprised of the following:solid platinum wall,liquid argon,and argon vapor.A few lay-ers of the liquid argon were placed on the nanoposts dec-orated solid surface.The nanoposts height was varied keep-ing the liquidfilm thickness constant to capture three scenar-ios:(i)liquid-film thickness is higher than the height of the nanoposts,(ii)liquid-film and nanoposts are of same height, and(iii)liquid-film thickness is less than the height of the nanoposts.The rest of the simulation box wasfilled with ar-gon vapor.The simulation was started from its initial config-uration,and once the equilibrium of the three phase system was established,the wall was suddenly heated to a higher temperature which resembles an ultrafast laser heating.Two different jump temperatures were selected:a few degrees above the boiling point to initiate normal evaporation and far above the critical point to initiate explosive boiling.Simula-tion results indicate nanostructures play a significant role in both cases:Argon responds very quickly for the nanostruc-tured surface,the transition from liquid to vapor becomes more gradual,and the evaporation rate increases with the nanoposts height.1IntroductionUnderstanding the phase transition phenomena of thin liq-uidfilms on a solid surface is of great importance for its A.K.M.M.Morshed·T.C.Paul·J.A.Khan( )Department of Mechanical Engineering,University of South Carolina,Columbia,SC29208,USAe-mail:khan@Fax:+1-803-7770106diverse practical applications such as thermal management of nanoelectronics[1],laser steam cleaning of a solid sur-face[2],colloidal deposition on a solid[3],laser surgery [4],etc.Molecular dynamics simulation is a powerful tool to capture the microscopic view of the nanoscale phenom-ena.Previously,molecular dynamics simulations have been carried out by a number of researchers[5–9]to understand the molecular behavior of the phase transition from liquid to vapor.Most of the researchers’focus was on the hetero-geneous phase transition of liquid from a molecular level flat solid surface.Phase transition from a solid surface can occur in a normal boiling mode or in an explosive boiling mode depending on the surface superheat temperature[10]. Although practically most of the solid surfaces have some molecular level unevenness and with the advent of ultrafine manufacturing technology,it is possible to precisely control the nanoscale structure on a solid surface but it has drawn very little attention of the researchers.In the macroscopic view,the molecular level roughened surface can be consid-ered as aflat surface but that roughness plays a very sig-nificant role in heterogeneous boiling[11]by changing the surface wettability[12]and reducing interfacial thermal re-sistance[13].In the present study,boiling both in normal and the ex-plosive mode over a nanostructured surface has been inves-tigated.Nanostructures heights were varied keeping liquid layers thickness constant.Three different initial configura-tions were used:nanostructure’s height higher than the liq-uidfilm thickness,nanostructure and liquid of same height, and nanostructure’s height less than the liquidfilm thick-ness.A nonequilibrium molecular dynamics(NEMD)sim-ulation was carried out using a simplified liquid and solid model to capture the microscopic view of the phase tran-sition and associated other phenomena of the above men-446A.K.M.M.Morshed et al.Fig.1Simulation domain and wall configurations:(a )initial configuration of the simulation domain for flat surface;(b )flat surface with the liquid argon layer;(c )4nanoposts position on the solid surface;(d )nanostructured surface-1;(e )nanostructured surface-2;(f )nanostructuredsurface-3tioned three cases.As far we know,no one has reported any investigation for these phenomena.2Simulation methodThe molecular system employed for this simulation con-sisted of a solid platinum wall,liquid argon layers,and an argon vapor in a cuboid of dimension 7.8165×32.1345×7.8165nm.Seven monolayers of the solid platinum wall were placed at the bottom of the simulation box.For the flat wall (surface-P),a total of 5,600platinum atoms were ar-ranged in a fcc (100)lattice structure corresponding to its density of 21.45×103kg/m 3.For the nanostructured sur-face,four cylindrical nanoposts each having a diameter of 1.7nm were placed on top of the solid wall.Three differ-ent heights (1.754nm (surface-1),3.19nm (surface-2),and 4.782nm (surface-3))were set for the nanoposts to obtain three different nanostructured wall configurations.Eleven monolayers of liquid argon atom were placed on top of the solid wall corresponding to a density of 1.367×103kg/m 3,and the rest of the space of the domain was filled with 200argon vapor atoms.Snapshots of the initial configuration of the domain and solid wall structure are presented in Fig.1.All the interactions between the atoms were mediated by the well-known Lenard–Jones (lj)12-6potential [14].φ(r)=4ε σr 12−σr6 Interaction potential’s length and energy parameters werecollected from [11]:σAr −Ar =0.3405nm,εAr −Ar =1.67×10−21J;σPt −Pt =0.2475nm,εPt −Pt =8.35×10−20J;σAr −Pt =0.294nm,εAr −Pt =1.65321×10−21J (only the hydrophilic case was considered).Although the Embedded Atom Method (EAM)potential is more appropriate for Pt-Pt interaction,nevertheless lj potential yields good results for a qualitative conclusion.To increase the computational efficiency,all the potentials were truncated at 3.5σAr −Ar .The equations of motions were integrated using velocity-varlet algorithm with 5fs time step.Periodic boundary con-ditions were imposed laterally (x and z direction).The bot-tom layer of the wall was kept fixed in order to avoid any migration of the sample [15]and a Langevin thermostat was applied to the next three layers of the wall;the adiabatic imaginary boundary condition was applied to the top of the simulation domain.To maintain the shape of the nanostruc-tures,the innermost core of the nanoposts were only allowedEffect of nanostructures on evaporation and explosive boiling of thin liquid films:a molecular dynamics447Fig.2Reduced density profile of argon on a flat surfaceto vibrate around their fixed lattice position by applying a spring force with an arbitrary spring constant of 57εσ2.The simulation was started from the initial configuration of the domain,the Langevin thermostat was switched on at 90K and the simulation continued for 1ns.Then the sim-ulation domain was allowed to equilibrate for another 1ns switching off the thermostat.During the equilibration pe-riod,the argon atom’s temperature,pressure,and density were monitored to check whether it was in the equilibrium state or not.For a flat surface,the temperature fluctuates around a mean value of 90.4K,and the density profile of argon atoms at the end of the equilibration period is pre-sented in pared with the phase diagram of the Lenard–Jones system [10,16],they are indeed in an equi-librium liquid-vapor system.At the end of the equilibrium period,to initiate phase transition,the Langevin thermo-stat was set to jump to 130K for the normal boiling case and 300K for the explosive boiling case.The upper wall (top three monolayers of the platinum wall)responds very quickly and it reaches equilibrium with the target tempera-ture within 50ps;the simulation was run for 4ns.All the simulations were performed using LAMMPS [17]and visu-alization was done using VMD [18].3Results and discussion 3.1High temperature caseFor the high temperature case,the bottom wall was set to jump from 90K to 300K,which is far away from the crit-ical temperature of the liquid and liquid argon goes into the fast phase transition or explosive boiling.Figure 3shows the temperature history of the wall and argon region.The wall responds very quickly to the temperature rise and reaches equilibrium with the target temperature within 50ps.Ini-tially,the argon region’s temperature also rises very sharply,and at some point it starts decreasing which indicates the separation of the argon layer from the wall surface.DueFig.3Temperature history of the argon and wall region for all the surface structures for the high temperaturecaseFig.4Pressure history of the simulation domain for high temperature caseto rapid rise of the wall temperature,liquid molecules ad-jacent to the solid wall exceed the critical point tempera-ture and vaporize while other layers are still in the liquid phase.The pressure of the vaporized liquid layer pushes the liquid above it and separates it from the solid wall.The low density vapor region inhibits the energy flow from the wall to the separated liquid and due to sudden expansion,the overall temperature of the liquid region falls temporar-ily.The solid–liquid interface area and interaction increases due to the presence of nanostructures.By the time liq-uid molecules adjacent to solid wall reaches the super-heated temperature (enough to initiate the separation),liq-uid molecules above that layer get more heated compared to the flat surface by getting energy from the nanostructures.As a result,the overall separation temperature increases due to presence of nanostructures as shown in Fig.3.It is also noticeable from Fig.3that the separation temperature in-creases sharply when the nanostructure’s height is equal or more than the liquid layer’s thickness.The plausible mech-anism of the fact is described later on in this section.The argon regions temperature starts increasing again,and for448A.K.M.M.Morshed et al.Fig.5Snapshots from the simulation (x –y plane)for the high temperature case.[(a )surface-P;(b )surface-1;(c )surface-2;(d )surface-3]the nanostructured surface-3,it reaches equilibrium around 4.5ns followed by the nanostructured surface-2,which reaches equilibrium around 5ns,and the other two surfaces do not reach equilibrium within the simulation time.Fig-ure 4shows the pressure history of the simulation domain.Due to volume constraint with the increase of temperature,the pressure also increases and it exactly follows the temper-ature trend of the simulation domain.Snapshots from the simulations are presented in Fig.5,which gives a molecular insight into the phase transition.At the end of the equilibration period (t =2ns),liquid argon shows a lower meniscus due to surface tension for surface-2and surface-3.For the flat surface,energy transfers only from the bottom and a temperature gradient is established in the liquid.At ∼150ps,liquid layers one or two layers above the solid surface vaporize first and push the liquidEffect of nanostructures on evaporation and explosive boiling of thin liquidfilms:a molecular dynamics449Fig.6Reduced density profile of the simulation domain for the high temperature case. ((a)2.15ns;(b)2.5ns)layer above it to separate from the solid as a large cluster of liquid as described earlier.For surface-1,vaporization and separation has been observed in the same fashion,but the separation starts at around100ps from the liquid layer just above the nanostructures.For nanostructured surface-2,the liquid layer is heated from the bottom and side as well;the temperature gradient in the liquid layer is less extensive than theflat surface.Due to heating,the liquid layer’s volume in-creases and it moves upward.Separation starts from the liq-uid just above the nanostructures.A small cluster of liquid moves upward and the rest of the liquid moves as an individ-ual atom or in a well dispersed tiny cluster.For the nanos-tructured surface-3,liquid escapes from the solid surface as an individual molecule or a well dispersed tiny cluster.Figure6shows the reduced density profile of the simula-tion domain.A cluster of liquid starts moving away from the flat surface at2.15ns;it already travels at small distances for nanostructured surface-1.A small density jump is ob-served for nanostructured surface-2,and for nanostructured surface-3two distinct density regions are observed but there is no density jump.At2.5ns,argon density becomes uni-form for the nanostructured sufaces-2and3but a large den-sity jump is observed for theflat surface and a small density jump is observed for the nanostructured surface-1.3.2Low temperature caseSimulation results for low temperature case,when the lower wall was set to jump from90K to130K are presented in this section.Figure7presents the temperature history of the argon region for all the surfaces.For the low temperature case,liquid argon molecules escapes into the vapor region from the top layer as an individual molecule and unlike the high temperature case,argon atoms for all the surfaces reach equilibrium.Due to the presence of nanostructures,interac-tion between the solid and liquid molecules increases as de-scribed earlier resulting in a faster energy transfer from the solid wall to the liquid molecules.Besides the average con-duction,distance of the liquid layer(from solid molecules to the top liquid layer)decreases for the nanostructures.As a result,the liquid response is quicker with the presence of nanostructures.At the end of the equilibration periodargon Fig.7History of argon temperature at different time steps for the low temperature caseatoms cover the nanostructures and create a meniscus for surface2and3,as discussed earlier.The average conduction distance of the liquid layers is almost the same for surfaces-2and3,and as a result they reach equilibrium at almost the same speedFigure8shows the reduced density profile of argon at dif-ferent time steps.A high density region is observed for all the surfaces near to the wall which indicates that a nonevap-orative liquidfilm exists.A few molecular layers of argon are also adsorbed on the nanostructures.As a result,the number of nonevaporative liquid molecules increases with the nanostructures.As the nonevaporativefilm thickness is equal to the height of the nanostructures length,the non-evaporative liquidfilm thickness is almost the same for the flat surface and surface-1.Figure9shows the net evapora-tion number for all the surfaces at different time steps.The evaporation rate was calculated by counting the change of argon molecules in the vapor regime for each5,000time steps.At the beginning,the evaporation rate increases al-most linearly for all the surfaces,and then it starts decreas-ing.The evaporation rate decreases earlier for surface-3fol-lowed by surface-2and surface-1,which implies that the evaporation rate is higher for long nanostructures.About 90%of the total evaporation is completed at5.25,4.4,3.875, and3.775ns for theflat surface,surface-1,surface-2,and surface-3,respectively.The evaporation rate for the initial450 A.K.M.M.Morshed et al.Fig.8Reduced density profile of argon at different time steps for the low temperature case. [(a)t=2ns;(b)2.5ns;(c)4.5ns;(d)6ns]Fig.9Net evaporation number as a function of timelinear zone is estimated at854and1248kg/m2s for aflat surface and nanostructured surface-1,respectively,for the nanostructured surface-2and surface-3;it is almost equal in the linear zone and is estimated at1392kg/m2s.4ConclusionThe phase transition of ultra-thin liquidfilms from a nanos-tructured surface when the nanostructure’s height is compa-rable to the liquidfilm thickness has been demonstrated us-ing molecular dynamics simulation.Main observations from the simulation results can be concluded as shown below: 1.The heat transfer rate from a solid surface to the liquidis enhanced for the nanostructured surface compared to theflat surface and the enhancement increases with the height of the nanostructures.2.In a fast phase transition when the liquid is suddenlyheated to a very high temperature,liquid molecules above the solid surface go into explosive boiling and a cluster of liquid is observed to move upward.The size of the cluster depends on the length of the nanostruc-tures.When the nanostructures length is more than the liquidfilm thickness,liquid molecules move as individ-ual molecules or a very tiny cluster instead of a large cluster of liquid.3.For the high temperature case,no nonevaporative thinliquidfilm is observed for any surfaces but the separation temperature is found to strongly depend on the length of the nanostructures.When the nanostructures length is equal or more than the liquidfilm thickness,the separa-tion temperature increases sharply.4.When the liquid layer is heated reasonably above theboiling point,evaporation starts from the top of the liquid surface and a nonevaporative thin liquid layer is observed for all the surface structures.5.The evaporation rate increases for the nanostructured sur-face compared to theflat surface.The evaporation rate does not vary significantly with the height of the nanos-tructures when it is equal or more than the liquid layer’s thickness.Acknowledgements This work was supported by the ESRDC con-sortium funded by the US Office of Naval Research under the Grant N00014-08-1-0080,and by the Department of Mechanical Engineer-ing at the University of South Carolina.Effect of nanostructures on evaporation and explosive boiling of thin liquidfilms:a molecular dynamics451References1.O.A.G.Kabov,E.Ya,D.V.Zaitsev,in Thermal and Thermome-chanical Phenomena in Electronic Systems2008,Orlando,FL, 28–31May20082.S.Georgiou,A.Koubenakis,Chem.Rev.103(2),349–394(2003)3.R.Bhardwaj,X.Fang,D.Attinger,New J.Phys.11(7),075020(2009)4.T.Juhasz,X.H.Hu,L.Turi,Z.Bor,Lasers Surg.Med.15(1),91–98(1994)5.P.Yi,D.Poulikakos,J.Walther,G.Yadigaroglu,Int.J.Heat MassTransf.45(10),2087–2100(2002)6.G.Nagayama,T.Tsuruta,P.Cheng,Int.J.Heat Mass Transf.49(23–24),4437–4443(2006)7. B.R.Novak,E.J.Maginn,M.J.McCready,J.Heat Transf.130(4),042411(2008)8.S.Maruyama,T.Kimura,in5th ASME–JSME Thermal Engineer-ing Joint Conference,San Diego,USA,19999.S.C.Maroo,J.N.Chung,J.Colloid Interface Sci.328(1),134–146(2008)10.X.Gu,H.M.Urbassek,Appl.Phys.B,Lasers Opt.81(5),675–679(2005)11.G.Nagayama,M.Kawagoe,A.Tokunaga,T.Tsuruta,Int.J.Therm.Sci.49(1),59–66(2010)12.G.Nagayama,S.Shi-iki,T.Tsuruta,Trans.Jpn.Soc.Mech.Eng.B73(728),1084–1091(2007)13.G.Nagayama,M.Kawagoe,T.Tsuruta,in MNC2007-21410,Ky-oto,Japan,2007,pp.1–1014.J.E.Lennard-Jones,A.F.Devonshire,Philos.Trans.R.Soc.Lond.Ser.A,Math.Phys.Sci.163(912),53–70(1937)15.Y.Dou,L.V.Zhigilei,N.Winograd,B.J.Garrison,J.Phys.Chem.105(12),2748–2755(2001)16. D.A.Kofke,J.Chem.Phys.98(5)(1993)17.S.J.Plimpton,put.Phys.117,1–19(1995)18.W.Humphrey,A.Dalke,K.Schulten,J.Mol.Graph.Model.14,33–38(1996)。

Package‘nephro’October16,2023Type PackageTitle Utilities for NephrologyVersion1.4Date2023-10-10Author Cristian Pattaro,Ryosuke Fujii,Janina HeroldMaintainer Cristian Pattaro<**************************>Description Set of functions to estimate kidney function and other phenotypes of interest in nephrol-ogy based on different biomechimal traits.License GPL(>=3)NeedsCompilation noRepository CRANDate/Publication2023-10-1621:40:02UTCR topics documented:nephro-package (2)CG (4)CKDEpi.creat (5)CKDEpi.creat.cys (6)CKDEpi.cys (7)CKDEpi_RF.creat (7)CKDEpi_RF.creat.cys (8)EKFC.creat (9)EKFC.cys (10)EKFC_SF.cys (10)FAS.creat (11)FAS.creat.cys (12)FAS.cys (13)MDRD4 (13)MDRD6 (14)Schwartz.Bedside (15)Stevens.creat.cys (16)1Stevens.cys1 (17)Stevens.cys2 (18)Virga (19)Index20 nephro-package Biostatistics utilities for nephrologyDescriptionThis package contains a set of tools for the estimation of kidney function.Kidney function is assessed by means of the Glomerular Filtration Rate(GFR),which can be estimated using different biomarkers.The most commonly used ones are serum or plasma creatinine and cystatin C.Included in the package are the following GFR estimating functions:the Modification of Diet in Re-nal Disease(MDRD)study equations based on four(MDRD4)or six(MDRD6)parameters(Levey1999;Levey2006);the CKD-Epi equations for serum creatinine with the race coefficient(CKDEpi.creat) and without the race coefficient(CKDEpi_RF.creat);the CKD-Epi equation for cystatin C(CKDEpi.cys);the CKD-Epi equation for the combination of creatinine and cystatin C with(CKDEpi.creat.cys) and without(CKDEpi_RF.creat.cys)the race coefficient(Inker2012;Inker2021);the three equa-tions proposed by Stevens2008based on cystatin C only(Stevens.cys1),age-and sex-weighted cystatin C(Stevens.cys2),and a combination of cystatin C and creatinine(Stevens.creat.cys);the classic Cockroft and Gault1976equation for creatinine clearance estimation(CG);the equation by Virga(2007)(Virga);the race-free equations developed by the European Kidney Function Con-sortium(EKFC)including sex and age based on serum creatinine(EKFC.creat)(Pottel2021)and based on serum cystatin C with(EKFC.cys)and without(EKFC_SF.cys)the sex coefficient(Pottel 2023);the full age spectrum(FAS)equations using serum creatinine(FAS.creat)(Pottel2016), cystatin C(FAS.cys),and their combination(FAS.creat.cys)(Pottel2017);the Schwartz bedside formula(Schwartz.Bedside)(Schwartz2009).A comparative description of several functions included in the initial version of the package can befound in Pattaro(2013).Extensive literature does exist that compares the methods described.DetailsPackage:nephroextType:PackageVersion: 1.4Date:2023-10-10License:GPLv3Author(s)Cristian Pattaro<**************************>,Ryosuke Fujii<***********************>,Janina Herold<***********************************.de>Acknowledgements:Max Plischke,Joao Sabino,Xiao Jiang for bug reporting;Alexander Tolios, Thomas Winkler for input on the software package;Andrew Srisuwananukorn for contributing scripts;the’nephro’user community for continuous feedback.ReferencesCiting this package:-Pattaro C,Riegler P,Stifter G,Modenese M,Minelli C,Pramstaller PP.Estimating the glomerular filtration rate in the general population using different equations:effects on classification and asso-ciation.Nephron Clin Pract2013;123(1-2):102-11.Formulas:-Cockroft DW,Gault MH.Prediction of creatinine clearance from serum creatinine.Nephron1976;16:31-41.-Inker LA,et al.Estimating glomerularfiltration rate from serum creatinine and cystatin C.N Engl J Med2012;367:20-9.-Inker LA,et al.New Creatinine-and Cystatin C-based Equations to Estimate GFR without Race.N Engl J Med2021;385:1737-1749.-Levey AS,et al.A more accurate method to estimate glomerularfiltration rate from serum crea-tinine:a new prediction equation.Modification of Diet in Renal Disease Study Group.Ann Intern Med.1999;130(6):461-70.-Levey AS,et ing standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerularfiltration rate.Ann Intern Med.2006;145:247-54.-Stevens LA,et al.Estimating GFR using serum cystatin C alone and in combination with serum creatinine:a pooled analysis of3,418individuals with CKD.Am J Kidney Dis2008;51:395-406.-Virga G,et al.A new equation for estimating renal function using age,body weight and serum creatinine.Nephron Clin Pract2007;105:c43-53.-Pottel H,et al.Development and Validation of a Modified Full Age Spectrum Creatinine-Based Equation to Estimate Glomerular Filtration Rate:A Cross-sectional Analysis of Pooled Data.Ann Intern Med2021;174:183-191.-Pottel H,et al.Cystatin C-Based Equation to Estimate GFR without the Inclusion of Race and Sex.N Engl J Med2023;388:333-343.-Pottel H,et al.An estimated glomerularfiltration rate equation for the full age spectrum.Nephrol Dial Transplant2016;31:798-806.-Pottel H,et al.Estimating glomerularfiltration rate for the full age spectrum from serum creati-nine and cystatin C Nephrol Dial Transplant2017;32:497-507.-Schwartz GJ,et al.New equations to estimate GFR in children with CKD.J Am Soc Nephrol 2009;20:629-637.On IDMS calibration:-Levey AS,et al.Expressing the Modification of Diet in Renal Disease Study equation for es-timating glomerularfiltration rate with standardized serum creatinine values.Clin Chem2007;53:766-72.-Matsushita K,et parison of risk prediction using the CKD-EPI equation and the MDRD study equation for estimated glomerularfiltration rate.J Am Med Assoc2012;307:1941-51.-Skali H,et al.Prognostic assessment of estimated glomerularfiltration rate by the new Chronic4CG Kidney Disease Epidemiology Collaboration equation in comparison with the Modification of Diet in Renal Disease Study equation.Am Heart J2011;162:548-54.Examples#Comparison between different equationscreat<-c(0.8,0.9,1.0,1.1,1.2,1.3)cyst<-c(1.1,0.95,1.1,1.0,1.3,1.2)sex<-c(1,1,1,0,0,0)age<-c(60,65,43,82,71,55)ethn<-round(runif(6))wt<-c(70,80,60,55,87,71)eGFR<-data.frame(creat,cyst)eGFR$MDRD4<-MDRD4(creat,sex,age,ethn, IDMS )eGFR$CKDEpi.creat<-CKDEpi.creat(creat,sex,age,ethn)eGFR$CKDEpi_RF.creat<-CKDEpi_RF.creat(creat,sex,age)eGFR$CKDEpi.cys<-CKDEpi.cys(cyst,sex,age)eGFR$CKDEpi.creat.cys<-CKDEpi.creat.cys(creat,cyst,sex,age,ethn)eGFR$CKDEpi_RF.creat.cys<-CKDEpi_RF.creat.cys(creat,cyst,sex,age)eGFR$Stevens.cys1<-Stevens.cys1(cyst)eGFR$Stevens.cys2<-Stevens.cys2(cyst,sex,age,ethn)eGFR$Stevens.creat.cys<-Stevens.creat.cys(creat,cyst,sex,age,ethn)eGFR$cg<-CG(creat,sex,age,wt)eGFR$virga<-Virga(creat,sex,age,wt)pairs(eGFR[,3:13])#For use with non-IDMS calibrated creatinine#several authors(see references)suggested#a5%creatinine adjustmentcreat<-c(0.8,0.9,1.0,1.1,1.2,1.3)sex<-c(1,1,1,0,0,0)age<-c(60,65,43,82,71,55)ethn<-round(runif(6))gfr<-CKDEpi.creat(0.95*creat,sex,age,ethn)CG Cockroft and Gault equationDescriptionCreatinine clearance is estimated with the Cockroft and Gault formula.UsageCG(creatinine,sex,age,wt)CKDEpi.creat5Argumentscreatinine Numeric vector with serum or plasma creatinine values in mg/dlsex Numeric vector with0for females and1for malesage Numeric vector with age in yearswt Numeric vector with weight in kgValueA numeric vector with eGFR values in ml/min/1.73m2.Author(s)Cristian PattaroReferencesCockroft DW,Gault MH.Prediction of creatinine clearance from serum creatinine.Nephron1976;16:31-41.See AlsoCKDEpi.creat,MDRD4,VirgaCKDEpi.creat CKD-EPI equation for serum creatinineDescriptionGFR is estimated with the CKD-EPI Study equation based on IDMS serum or plasma creatinine.UsageCKDEpi.creat(creatinine,sex,age,ethnicity)Argumentscreatinine Numeric vector with serum or plasma creatinine values in mg/dlsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsethnicity Numeric vector with0for non-Black and1for Black individualsValueA numeric vector with eGFR values in ml/min/1.73m2.6CKDEpi.creat.cysAuthor(s)Cristian PattaroReferencesInker LA,et al.Estimating glomerularfiltration rate from serum creatinine and cystatin C.N Engl J Med2012;367:20-29.See AlsoCKDEpi.creat.cys,CKDEpi.cys,CKDEpi_RF.creatCKDEpi.creat.cys CKD-EPI equation for creatinine and cystatin CDescriptionCKD-EPI equation to estimate GFR based on a combination of creatinine and cystatin CUsageCKDEpi.creat.cys(creatinine,cystatin,sex,age,ethnicity)Argumentscreatinine Numeric vector with serum or plasma creatinine values in mg/dlcystatin Numeric vector with serum or plasma cystatin C values in mg/dlsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsethnicity Numeric vector with0for non-Black and1for Black individualsValueA numeric vector with eGFR values in ml/min/1.73m2.Author(s)Cristian PattaroReferencesInker LA,et al.Estimating glomerularfiltration rate from serum creatinine and cystatin C.N Engl J Med2012;367:20-29.See AlsoCKDEpi.creat,CKDEpi.cys,CKDEpi_RF.creat.cysCKDEpi.cys7 CKDEpi.cys CKD-EPI equation for cystatin CDescriptionGFR is estimated with the CKD-EPI equation for cystatin C proposed by Inker et al.,N Engl J Med 2012UsageCKDEpi.cys(cystatin,sex,age)Argumentscystatin Numeric vector with serum or plasma cystatin C values in mg/lsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsValueThe function returns a numeric vector with eGFR values in ml/min/1.73m2.Author(s)Cristian PattaroReferencesInker LA,et al.Estimating glomerularfiltration rate from serum creatinine and cystatin C.N Engl J Med2012;367:20-29.See AlsoCKDEpi.creat,CKDEpi.creat.cysCKDEpi_RF.creat Race-free CKD-EPI equation for serum creatinineDescriptionGFR is estimated with the CKD-EPI Study equation based on serum creatinine without the ethnicity coefficient.UsageCKDEpi_RF.creat(creatinine,sex,age)8CKDEpi_RF.creat.cysArgumentscreatinine Numeric vector with serum or plasma creatinine values in mg/dlsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsValueA numeric vector with eGFR values in ml/min/1.73m2.Author(s)Ryosuke FujiiReferencesInker LA,et al.New creatinine-and cystatin C-based equations to estimate GFR without race.N Engl J Med2021;385:1737-1749.See AlsoCKDEpi.creat,CKDEpi.creat.cys,CKDEpi.cys,CKDEpi_RF.creat.cysCKDEpi_RF.creat.cys Race-free CKD-EPI equation for serum creatinine and cystatin CDescriptionCKD-EPI equation to estimate GFR based on a combination of creatinine and cystatin C without the ethnicity coefficientUsageCKDEpi_RF.creat.cys(creatinine,cystatin,sex,age)Argumentscreatinine Numeric vector with serum or plasma creatinine values in mg/dlcystatin Numeric vector with serum or plasma cystatin C values in mg/lsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsValueA numeric vector with eGFR values in ml/min/1.73m2.EKFC.creat9Author(s)Ryosuke FujiiReferencesInker LA,et al.New creatinine-and cystatin C-based equations to estimate GFR without race.N Engl J Med2021;385:1737-1749.See AlsoCKDEpi.creat,CKDEpi.creat.cys,CKDEpi.cys,CKDEpi_RF.creatEKFC.creat EKFC equation for serum creatinineDescriptionEKFC equation for serum creatinine modified from FAS equationUsageEKFC.creat(creatinine,sex,age)Argumentscreatinine Numeric vector with serum or plasma creatinine values in mg/dlsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsValueThe function returns a numeric vector with eGFR values in ml/min/1.73m2.Author(s)Janina HeroldReferencesPottel,H,et al.Development and Validation of a Modified Full Age Spectrum Creatinine-Based Equation to Estimate Glomerular Filtration Rate:A Cross-sectional Analysis of Pooled Data.N Engl J Med2021;174:183-191.See AlsoEKFC.cys,CKDEpi.creat10EKFC_SF.cys EKFC.cys EKFC equation for cystatin CDescriptionEKFC equation for cystatin C that includes the sex coefficient,as proposed by Pottel et al.,N Engl J Med2023UsageEKFC.cys(cystatin,sex,age)Argumentscystatin Numeric vector with serum or plasma cystatin C values in mg/lsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsValueThe function returns a numeric vector with eGFR values in ml/min/1.73m2.Author(s)Janina HeroldReferencesPottel,H,et al.Cystatin C-Based Equation to Estimate GFR without the Inclusion of Race and Sex.N Engl J Med2023;388:333-343.See AlsoEKFC_SF.cysEKFC_SF.cys Sex-free EKFC equation for cystatin CDescriptionEKFC equation for cystatin C without the sex coefficient as proposed by Pottel et al.,N Engl J Med 2023UsageEKFC_SF.cys(cystatin,age)FAS.creat11Argumentscystatin Numeric vector with serum or plasma cystatin C values in mg/lage Numeric vector with age in yearsValueThe function returns a numeric vector with eGFR values in ml/min/1.73m2.Author(s)Janina HeroldReferencesPottel,H,et al.Cystatin C-Based Equation to Estimate GFR without the Inclusion of Race and Sex.N Engl J Med2023;388:333-343.See AlsoEKFC.cysFAS.creat FAS equation for serum creatinineDescriptionFull age spectrum(FAS)equation to estimate GFR based on serum creatinineUsageFAS.creat(creatinine,sex,age)Argumentscreatinine Numeric vector with serum or plasma creatinine values in mg/dlsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsValueA numeric vector with eGFR values in ml/min/1.73m2.Author(s)Janina Herold12FAS.creat.cys ReferencesPottel,H.,et al.An estimated glomerularfiltration rate equation for the full age spectrum.Nephrol Dial Transplant.2016;5:798-806.See AlsoFAS.creat.cys,FAS.cysFAS.creat.cys FAS equation for creatinine and cystatin CDescriptionFull age spectrum(FAS)GFR estimation based on serum creatinine and cystatin CUsageFAS.creat.cys(creatinine,cystatin,sex,age)Argumentscreatinine Numeric vector with serum or plasma creatinine values in mg/dlcystatin Numeric vector with serum or plasma cystatin C values in mg/lsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsValueThe function returns a numeric vector with eGFR values in ml/min/1.73m2.Author(s)Janina HeroldReferencesPottel,H.,et al.An estimated glomerularfiltration rate equation for the full age spectrum from serum creatinine and cystatin C.Nephrol Dial Transplant.2017;32:497-507.See AlsoFAS.cysFAS.cys13 FAS.cys FAS equation for cystatin CDescriptionFull age spectrum(FAS)GFR estimation based on cystatin CUsageFAS.cys(cystatin,sex,age)Argumentscystatin Numeric vector with serum or plasma cystatin values in mg/lsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsValueA numeric vector with eGFR values in ml/min/1.73m2.Author(s)Janina HeroldReferencesPottel,H.,et al.An estimated glomerularfiltration rate equation for the full age spectrum from serum creatinine and cystatin C.Nephrol Dial Transplant.2017;32:497-507.See AlsoFAS.creat,FAS.creat.cysMDRD4Four-parameter MDRD study equationDescriptionGFR is estimated with the4-parameter Modification of Diet in Renal Disease(MDRD)study equa-tion.UsageMDRD4(creatinine,sex,age,ethnicity,method="IDMS")14MDRD6 Argumentscreatinine Numeric vector with serum or plasma creatinine values in mg/dlsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsethnicity Numeric vector with0for non-Black and1for Black individualsmethod Defaults is’IDMS’for IDMS-traceable creatinine;write’other’if not IDMSValueA numeric vector with eGFR values in ml/min/1.73m2.Author(s)Cristian PattaroReferencesLevey AS,et al.A more accurate method to estimate glomerularfiltration rate from serum creati-nine:a new prediction equation.Modification of Diet in Renal Disease Study Group.Ann Intern Med.1999;130(6):461-70.Levey AS,et ing standardized serum creatinine values in the modification of diet in renal dis-ease study equation for estimating glomerularfiltration rate.Ann Intern Med.2006;145:247-254. See AlsoCKDEpi.creat,MDRD6,CGMDRD6Six-parameter MDRD study equationDescriptionGFR is estimated with the6-parameter Modification of Diet in Renal Disease(MDRD)study equa-tion.UsageMDRD6(creatinine,sex,age,albumin,BUN,ethnicity,method= IDMS )Schwartz.Bedside15Argumentscreatinine Numeric vector with serum or plasma creatinine values in mg/dlsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsalbumin Numeric vector with serum or plasma albumin in g/dlBUN Numeric vector with blood urea nitrogen levels in mg/dlethnicity Numeric vector with0for non-Black and1for Black individualsmethod Defaults is’IDMS’for IDMS-traceable creatinine;write’other’if not IDMS ValueA numeric vector with eGFR values in ml/min/1.73m2.Author(s)Cristian PattaroReferencesLevey AS,et al.A more accurate method to estimate glomerularfiltration rate from serum creati-nine:a new prediction equation.Modification of Diet in Renal Disease Study Group.Ann Intern Med.1999;130(6):461-70.Levey AS,et ing standardized serum creatinine values in the modification of diet in renal dis-ease study equation for estimating glomerularfiltration rate.Ann Intern Med.2006;145:247-254. See AlsoMDRD4Schwartz.Bedside Bedside IDMS-traceable Schwartz GFR Calculator for ChildrenDescriptionGFR is estimated with the Bedside Schwartz equation for Children based on IDMS serum or plasma creatinine.This equation is valid in the1-17years age range.UsageSchwartz.Bedside(creatinine,ht,age)16Stevens.creat.cysArgumentscreatinine Numeric vector with serum or plasma creatinine values in mg/dlht Numeric vector with height in cmage Numeric vector with age in yearsValueA numeric vector with eGFR values in ml/min/1.73m2.Author(s)Andrew SrisuwananukornReferencesSchwartz GJ,et al.New equations to estimate GFR in children with CKD.J Am Soc Nephrol2009;20:629-637.Stevens.creat.cys Stevens’formula for a combination of serum creatinine and cystatin CDescriptionGFR estimation using the3rd formula proposed by Stevens et al.(Am J Kidney Dis2008),which combines creatinine and cystatin CUsageStevens.creat.cys(creatinine,cystatin,sex,age,ethnicity)Argumentscreatinine Numeric vector with serum or plasma creatinine values in mg/dlcystatin Numeric vector with serum or plasma cystatin C values in mg/lsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsethnicity Numeric vector with0for non-Black and1for Black individualsValueThe function returns a numeric vector with eGFR values in ml/min/1.73m2.Author(s)Cristian PattaroReferencesStevens LA,et al.Estimating GFR using serum cystatin C alone and in combination with serum creatinine:a pooled analysis of3,418individuals with CKD.Am J Kidney Dis2008;51:395-406. See AlsoCKDEpi.creat.cysStevens.cys1GFR estimation using serum cystatin CDescriptionGFR is estimated with the1st formula proposed by Stevens et al.(Am J Kidney Dis2008),i.e.:asa simple transformation of cystatin C,without using any other informationUsageStevens.cys1(cystatin)Argumentscystatin Numeric vector with serum or plasma cystatin C values in mg/lValueA numeric vector with eGFR values in ml/min/1.73m2.Author(s)Cristian PattaroReferencesStevens LA,et al.Estimating GFR using serum cystatin C alone and in combination with serum creatinine:a pooled analysis of3,418individuals with CKD.Am J Kidney Dis2008;51:395-406. See AlsoStevens.cys2,Stevens.creat.cys,CKDEpi.cysStevens.cys2Stevens’formula for serum cystatin C,age,and sexDescriptionGFR is estimated with the2nd formula proposed by Stevens et al.(Am J Kidney Dis2008),where cystatin C is weighted by sex and ageUsageStevens.cys2(cystatin,sex,age,ethnicity)Argumentscystatin Numeric vector with serum or plasma cystatin C values in mg/lsex Numeric vector with0for females and1for malesage Numeric vector with age in yearsethnicity Numeric vector with0for non-Black and1for Black individualsValueA numeric vector with eGFR values in ml/min/1.73m2.Author(s)Cristian PattaroReferencesStevens LA,et al.Estimating GFR using serum cystatin C alone and in combination with serum creatinine:a pooled analysis of3,418individuals with CKD.Am J Kidney Dis2008;51:395-406. See AlsoStevens.cys1,Stevens.creat.cys,CKDEpi.cysVirga19 Virga Virga’s formulaDescriptionVirga’s formula is based on serum creatinine,sex,age,and body weight.UsageVirga(creatinine,sex,age,wt)Argumentscreatinine Numeric vector with serum or plasma creatinine values in mg/dlsex Numeric0/1vector:0for females,1for malesage Numeric vector with age in yearswt Numeric vector with weight in kgValueA numeric vector with eGFR values in ml/min/1.73m^2Author(s)Cristian PattaroReferencesVirga G,et al.A new equation for estimating renal function using age,body weight and serum creatinine.Nephron Clin Pract2007;105:c43-53.See AlsoCG,MDRD4Index∗CKD-EPICKDEpi.creat,5CKDEpi.creat.cys,6CKDEpi.cys,7CKDEpi_RF.creat,7CKDEpi_RF.creat.cys,8nephro-package,2∗EFKFEKFC.creat,9EKFC.cys,10EKFC_SF.cys,10FAS.creat.cys,12∗EKFCnephro-package,2∗FASFAS.creat,11FAS.cys,13nephro-package,2∗GFRCKDEpi.cys,7EKFC.creat,9EKFC.cys,10EKFC_SF.cys,10FAS.creat.cys,12nephro-package,2Stevens.creat.cys,16Stevens.cys1,17Stevens.cys2,18∗MDRDCG,4MDRD4,13MDRD6,14nephro-package,2∗PottelFAS.creat,11FAS.cys,13∗SchwartzSchwartz.Bedside,15∗VirgaVirga,19∗albuminMDRD6,14∗blood urea nitrogenMDRD6,14∗clearanceCG,4Virga,19∗creatinineCG,4CKDEpi.creat,5CKDEpi.creat.cys,6CKDEpi.cys,7CKDEpi_RF.creat,7CKDEpi_RF.creat.cys,8EKFC.creat,9FAS.creat,11MDRD4,13MDRD6,14nephro-package,2Schwartz.Bedside,15Stevens.creat.cys,16Virga,19∗cystatinCKDEpi.creat.cys,6CKDEpi.cys,7CKDEpi_RF.creat.cys,8EKFC.cys,10EKFC_SF.cys,10FAS.creat.cys,12FAS.cys,13nephro-package,2Stevens.creat.cys,16Stevens.cys1,17Stevens.cys2,18∗kidney functionnephro-package,2∗sexEKFC.cys,1020INDEX21 CG,2,4,14,19CKDEpi.creat,2,5,5,6–9,14CKDEpi.creat.cys,2,6,6,7–9,17CKDEpi.cys,2,6,7,8,9,17,18CKDEpi_RF.creat,2,6,7,9CKDEpi_RF.creat.cys,2,6,8,8EKFC.creat,2,9EKFC.cys,2,9,10,11EKFC_SF.cys,2,10,10FAS.creat,2,11,13FAS.creat.cys,2,12,12,13FAS.cys,2,12,13MDRD4,2,5,13,15,19MDRD6,2,14,14nephro(nephro-package),2nephro-package,2Schwartz.Bedside,2,15Stevens.creat.cys,2,16,17,18Stevens.cys1,2,17,18Stevens.cys2,2,17,18Virga,2,5,19。

13Machine Vision and ApplicationsISSN 0932-8092Machine Vision and Applications DOI 10.1007/s00138-016-0750-1Improved $${{\ell }^{1}}$$ ℓ 1 -tracker using robust PCA and random projectionDongjing Shan & Zhang ChaoYour article is protected by copyright andall rights are held exclusively by Springer-Verlag Berlin Heidelberg. This e-offprint isfor personal use only and shall not be self-archived in electronic repositories. If you wish to self-archive your article, please use the accepted manuscript version for posting on your own website. You may further deposit the accepted manuscript version in any repository, provided it is only made publicly available 12 months after official publication or later and provided acknowledgement is given to the original source of publicationand a link is inserted to the published article on Springer's website. The link must be accompanied by the following text: "The final publication is available at ”.13Improved 1-tracker using robust PCA and random projection Dongjing Shan1·Zhang Chao1Received:5April2015/Revised:20December2015/Accepted:4January2016©Springer-Verlag Berlin Heidelberg2016Abstract In this paper,we propose an improved 1-tracker in a particlefilter framework using robust principal compo-nent analysis(robust PCA)and random projection.Atfirst we redesign the template set and its update scheme.Three tar-get templates and several background templates combined with the trivial templates are used to represent the candi-date images sparsely.Onefixed target template is generated from the image patch in thefirst frame.The other two are dynamic target templates,called stable target template,and fast changing one used for long time and short time,respec-tively.Robust PCA is used to generate and update the stable target template,and fast changing target template is initial-ized by the stable one at certain times.The background templates are used to strengthen the ability of distinguishing background and foreground.Then,the large set of Haar-like features are extracted and compressively sensed with a very sparse measurement matrix for the 1-tracker framework. The compressive sensing theories ensure that the sensed features preserve almost all the information of the origi-nal features.Our proposed method is more robust than the original 1-method.Experiments have been done on numer-ous sequences to demonstrate the better performance of our improved tracker.Keywords 1-minimization·Sparse representation·Robust PCA·Random projection·Particlefilter·Visual trackingB Dongjing Shanshandongjing@1Key Laboratory of Machine Perception(MOE),Peking University,Beijing100871,China 1IntroductionSparse representation[6]is an emergingfield that provides a framework for reconstructing a signal(e.g.,an image)using sub-Nyquist sampling rates.Specifically,a signal can be reconstructed from a small set of random projections,pro-vided that the signal is sparse in some basis,e.g.,wavelets. The idea of sparse representation has recently been exploited for a variety of problems in computer vision,such as back-ground subtraction[9],media recovery[7],face recognition [17],object tracking[8,12],and so on.In visual tracking, different methods also have been developed based on sparse representation.In[20],tracking is formulated as a structured multi-task sparse learning problem,where particle represen-tations are learned jointly using the APG method,in[15], an online tracking algorithm is designed based on sparse prototype representation,and in[3],based on[12]a new minimization model for computing the sparse representation of the target is proposed and it is accelerated using APG.Our method is based on the 1-tracker[12];in the orig-inal algorithm,sparse representation is employed for visual tracking with the intuition that the appearance of a tracked object can be sparsely represented by its appearances in pre-vious frames.Specifically,tofind the tracking target in a new frame,each target candidate in the particlefilter framework is sparsely represented in the space spanned by target templates and trivial templates.The sparseness is achieved by solving a minimization problem which can be done efficiently through convex optimization.Then the candidate which has the small-est projection error with its sparse representation is chosen as the tracking target.In our work,robust PCA[16]and random projection[19] are exploited to strengthen the algorithm.First,only three target templates and several background templates are added into the template set.One of the target templates is afixed123S.Dongjing,Z.Chao template created from a manually selected target in thefirstframe;the second is a dynamic template updated via robustPCA,which constructs a stable appearance model for long-time tracking;The third is also a dynamic template whichis frequently reinitialized based on the stable template andupdated rapidly to reflect the fast appearance change of thetarget.The background templates consist of several back-ground image patches cropped from the background regionsof the former frames.They can strengthen the algorithm’sability of distinguishing the background and foreground anddecrease the probability of drifting.Obviously,combinedwith the trivial templates,these templates can be used torepresent the candidate image patches sparsely.So,the can-didate with the minimum distance to its linear combinationcorresponding to only the target templates will be adoptedas the tracking target.In other words,the candidate with theminimum projection error on the target templates will be thefittest result.Second,we extract reduced-dimensional Haar-like fea-tures for the framework of 1-tracker.The vector of large setof haar-like features extracted from the image patch is highdimensional and should be compressively sensed to reduce itsdimension.We adopt random projection theory to cope withthis problem.A very sparse measurement matrix that satis-fies the restricted isometry property(RIP)will make sure thesensed feature vector preserves almost all the original infor-mation.The reduced dimensional feature vector will be usedto describe all the templates(except for the trivial templates)and the candidate image patches.2ParticlefilterThe particlefilter is a sequential Monte Carlo algorithm[13],i.e.a sampling method for approximating a distribution ofstate variables characterizing a dynamic system.Concretely,given a set of observations z0:t,we aim to estimate the valueof the hidden state variable x ing Bayes theorem,we havethe familiar result:P(x t|z0:t)=αP(z t|x t)P(x t−1|z0:t−1)P(x t|x t−1)d x t−1(1)Then,P(x t−1|z0:t−1)can be represented using a collection of N weighted samples or particles:{x(i)t−1,ω(i)t−1}N i=1,where ω(i)t−1is the weight of particle x(i)t−1.A particle representation of this density isP(x t−1|z0:t−1)≈Ni=1ω(i)t−1δx t−1−x(i)t−1(2)As a result,Eq.2can be approximated as follows:P(x t|z0:t)≈αP(z t|x t)Ni=1ω(i)t−1Px t|x(i)t−1(3)Apparently,the formulation(3)is governed by the obser-vation model P(z t|x t),where we estimate the likelihood of x t observing z t and the dynamic model between two states P(x t|x t−1).Dynamical model The state variable x t is modeled by six parameters of the affine transformation.x t= (α1,α2,α3,α4,t x,t y),where{α1,α2,α3,α4}are the deformation parame-ters and(t x,t y)are the2D translation parameters.We crop the region of candidate z t from the frame and normalize it to be the same size as the initial bounding box.Each parameter in x t is modeled independently by a Gaussian distribution around its counterpart in x t−1.Observation model the observation model P(z t|x t)reflects the similarity between a target candidate and the target tem-plate.In this paper,the observation model is formulated by the distance from the target candidate to its linear combina-tion of the target templates.3 1-tracker using robust PCA and random projection3.1Reduced dimensional featureIn our tracker,all the templates and candidates are rep-resented as normalized feature vectors.We use reduced dimensional haar-like feature[19]in the sparse represen-tation framework.As depicted in Fig.1,high-dimensional feature vectors(multi-scale haar-like feature:x∈R m and its dimensionality is typically in the order of106to1010) are extracted from the candidate image patches,R∈R n×m is the random projection matrix(n<<m),v∈R n is the reduced dimensional vector;we havev=Rx(4) The matrix R is designed as follows:r i,j=√s×⎧⎨⎩1with probability12s0with probability1−1s−1with probability12s(5)Fig.1Framework of extracting reduced dimensional Haar-like fea-tures123Improved 1-tracker using robust PCA...where s =O (m /log m )and n ≥κβlog (m /β)the matrix R will satisfy restricted isometry property (RIP)[1,4].In our work we set s =m /6and n =80;then R is sparse and satisfies RIP and can save a lot of computational effort and preserve almost all of the information.3.2Target templatesThree target templates are used for sparse representation (see Fig.2).One is called fixed target template (t f ),which is man-ually selected from the first frame.The second one is a stabletarget template (represented as t s d),which is updated dynami-cally using robust PCA [16]so as to construct a stable model for tracking process.First,a data set (D 0)is established based on the tracking results in the former n 0(10used here)frames.For the similarity of the tracking results,a low-rank matrix can be recovered from the data set by removing the gross corruption and even outliers:min A ,E||A 0||∗+λ||E 0||1,subject to D 0=A 0+E 0,(6)where D 0is a matrix with each column being a reduceddimensional feature vector from one tracking result which has been normalized.||.||∗means the nuclear norm of a matrix,||.||1denotes the sum of the absolute values of matrix entries,and λis a positive weighting parameter.We use the augmented lagrange multiplier method for exact recovery of corrupted low-rank matrix [5]to solve this problem.The resulting E 0is error matrix and A 0is the purified data matrix.And then,when the next n +(set 5here)tracking results are provided (denoted as D +),they should be used to update the data matrix D 0.So,[D 0D +]is purified by robust PCA,and we getmin[ A0A +],[ E 0E +]||[ A0A +]||∗+λ||[ E 0E +]||1,(7)subject to [D 0D +]=[ A0A +]+[ E 0E +],where [ A0A +]denotes the new recovered purified matrix and [ E0E +]is the new error matrix.Two things should be done if min i||e +i ||2≤max j||e 0j ||2+ ( makes the target image with certain distortion can be updated into the tem-plate,and is set 0.05here).The first one,the j th columnofFig.2The template set consists of the target templates,the backgroundtemplates and the trivial templatesmatrix D 0,can be replaced by i th column of matrix D +;the updated D 0will be used circularly when the next n +trackingresults come.The other one,the stable target template t s d,can be formulated as follows:t s d =f n f n +m t s 0d +m f n +mi m ,(8)where t s 0d denotes the previous target template and t s dmeans the current one.n denotes the efficient size of frames having been processed and at each update be computed as n ←f n +m ,m denotes the number of frames newly come,which is set to 1in our tracker,i m denotes the mean value of the target images in the m frames,i.e.i th column of A +(a +i )corresponding to the column e +i with minimum 2-norm in our implementation,f is a forgetting factor (set to 0.95here)that reduces the effect of previously tracked target images and keeps the template more consistent with the changing target.As for the third target template (named fast changing tem-plate t fd ),it is designed to be updated by each frame according to the formulation (8)and reinitialized by the stable template t s d at certain times.tf dis updated quickly so as to capture the fast appearance change of the target;every new tracked target is used to compute a new template [m in the formulation (8)is also set to 1here];obviously,it needs to be reinitialized frequently because inaccurate or even wrong tracking results may be included for consideration.Fortunately,the stabletemplate t s d is complementary with t f d;once t s d is updated,t f d should be reinitialized using t s d immediately and updated afterwards independently.All in all,as the stable templateis more reliable,it can set new start points for t fd which is keener in short time tracking.3.3Background templatesAs depicted in Fig.3,background templates are combinedinto the template set.First,we should construct a background image (i bg )using gaussian mixture model.Then,we use the former 5frames of the current tracking frame to create 5background frames (b i ),directly;we replace the foreground of each frame (b i )with a background patch cropped from the constructed background image (i bg ).At last,5background templates can be cropped from these frames and inserted into the background template set.As shown in Fig.3,for any can-didate y ,we select the corresponding image patches from the background frames according to the coordinate of y .Those patches will be inserted into background template set and apparently background templates change for different can-didate y .Background templates can improve the robustness of the tracker as the introduce of them can make the back-ground more distinguishable.We will explain how it works in Sect.3.4.123S.Dongjing,Z.ChaoFig.3Framework of the proposed method3.4Framework of sparse representationObviously,a tracking target lies approximately in the lowdimensional subspace[18]spanned by the training images ofthe target or the tracked images extracted from the precedingframes,the imprecision is caused by the changing appearanceof the target in the tracking process.Given target template setT1=[t f,t s d,t f d]∈R l×3(l>>3),containing three targettemplates which are reduced dimensional feature vector,andthe background templates T2=[t1b,t2b,...,t n b],a trackingresult y∈R l can be represented approximately as a linearcombination of the two target templates as follows:y≈T a=(T1∪T2)a=a1t f+a2t s d+a3t f d+a4t1b+···+a n+3t n b;(9)we use the trivial templates I=[i1,i2,...,i l]∈R l×l andimpose nonnegativity constraints on the coefficients[10]toformulate the Eq.(9)as follows:y=[T,I,−I]⎡⎣ae+e−⎤⎦=Bc,s.t.c≥0,(10)where e+∈R l,e−∈R l are called a positive trivial coefficient vector and a negative trivial coefficient vector, respectively,and c is a non-negative coefficient vector.We exploit the compressibility in the transform domain by solv-ing the problem(10)as an 1-regularized least squares problem,which is known to typically yield sparse solutions [17].min||Bc−y||22+λ||c||1,s.t.c≥0,(11) where||.||1and||.||2denote the 1and 2norms,respectively, c is the sparse coefficients,and we record the coefficients in c corresponding to the target templates as“c t”.We use the same method(interior-point method[12])as 1-tracker to solve Eq.11;it uses the preconditioned conjugate gradients (PCG)algorithm to compute the search direction,and the number of PCG iterations depends on the value ofλ,we set λto0.1here,and in the experiments we found that it can keep a balance between speed and accuracy.Then the process offinding the target is as fellows:∧y=arg max y∈γj=1,...,dN((y−T1c t)(j);0,δ2),(12)whereγ={y1,y2,...,y n p}are the n p target candidates, and T1is the target template set.The candidate with the min-imum distance will be chosen as the tracking target.Now we will explain why the background templates work for this algorithm.Obviously,if a candidate is not the fore-ground(background)but approximates the foreground in appearance,the original tracker[11]will probably drift to this candidate.Nevertheless,for our tracker,the coefficients on the target templates will be still sparse because the back-ground templates are more likely to be used to represent the candidate(background candidate).So,the distance from the candidate to its target templates combination will be still large.Therefore,it is easy to distinguish between foreground and background candidate similar to the target.4ExperimentsWe compare our tracker with IPCA[14], 1-tracker[12], TLD[11],MIL[2],MTT[20],CT[19],L1APG[3],and SP [15];also WOTRPCA is used for comparison,which is our tracker without the robust PCA,i.e.in the target template set of our tracker onlyfixed and fast changing templates are used. All the experiments are done with a computer of3.1GHz i5 CPU and4GB memory.Our method is implemented in MAT-LAB and takes about0.51s to process one frame while300123Improved 1-tracker using robustPCA...Fig.4CarFig.5DavidFig.6Girlparticles are used.We expect to achieve a realtime tracker by using the APG method in our future work,which is declared in the Conclusion.Quantitative results are displayed in Figs.4,5,6,7,8and Fig.9where center distance of the tracking rectangle to the ground truth is computed everyfive frames(as the ground truth is labeled at intervals offive frames).Mean value of the center distance is shown in the upper-left box of each figure.Background clutter(e.g.Fig.4),seriousocclusion Fig.7FaceoccFig.8Faceocc2Fig.9Faceocc2(e.g.Figs.7,8),appearance and illumination changing(e.g. Figs.5,9)are contained in these cases.Our tracker can cope with the difficulties while other trackers lose the object in some cases.Table1illustrates the percentage of frames correctly trac-ked with regard to seven publicly available sequences,where the bold values indicatefirst ranking and the italic values indicate second ranking.Given the detected bounding box ROI D and the ground truth bounding box ROI GT,the overlap123S.Dongjing,Z.ChaoTable1The percentage of frames correctly trackedSequences IPCA 1-tracker TLD MIL MTT CT L1APG SP WOTRPCA Our trackerGirl_occ[12]0.660.970.640.710.810.730.830.860.820.98 David[14]0.900.54 1.00.700.530.990.780.840.860.95Sylv[2]0.550.410.280.780.800.310.540.510.680.76 Faceocc1[2]0.990.98 1.00.930.940.80 1.00.910.97 1.0 Faceocc2[2]0.800.690.830.960.850.790.610.840.720.85Car[14] 1.00.600.630.410.970.560.980.610.95 1.0 Walker[10]0.670.670.830.460.820.710.900.860.840.92Fig.10The tracking results of“car”over frames#238,#250,#251,#253,#274,#290.Row1 1-tracker.Row2Our improvedtrackerFig.11The tracking results of“girl”[12]over frames of#56,#84,#97,#115,#122,#128.Row1 1-tracker.Row2Our improvedtrackerFig.12The tracking results of“sylv”over frames of#160,#232,#381,#509,#571,#600.Row1 1-tracker.Row2Our improved trackerscore is evaluated as follows:score=area(ROI D∩ROI G T)area(ROI D∪ROI G T).(13)If the score is greater than0.5,the tracking result will be labeled as right.Figures10,11,12show some picture results of 1-tracker and our improved tracker.In Fig.10, 1-tracker loses the car at frame#251for the background cluster while our improved one succeeds.In Fig.11, 1-tracker loses the object because of its large appearance change,our tracker can recapture the person for our updating scheme of the target templates.In Fig.12, 1-tracker cannot track the doll accurately while our tracker accomplish the task.5ConclusionIn this paper,we propose an improved tracking method which is more robust than the original 1-tracker when coping with challenging cases;robust PCA and random projection are used here to improve the performance.But some problems remain to be solved.For example,we can exploit the sparse-ness of the trivial coefficients more deeply to implement a more reliable fast changing template,and when the tar-get disappears,a modified dynamical model in particlefilter framework or trajectory analysis can be applied to make the tracker more stable in refinding the target.Additionally,the speed of our improved tracker can be accelerated signifi-cantly by using APG method(as declared in MTT[20],L∗11 tracker is120times faster than the 1-tracker as APG method is adopted to solve the optimization problem).We expect the issues can be addressed in our future work. References1.Achlioptas,D.:Database-friendly random projections:Johnson-lindenstrauss with binary put.Syst.Sci.66(4),671–687(2003)2.Babenko,B.,Yang,M.H.,Belongie,S.:Visual tracking with onlinemultiple instance learning.In:IEEE Conference on CVPR2009, pp.983–990(2009).doi:10.1109/CVPR.2009.52067373.Bao,C.,Wu,Y.,Ling,H.,Ji,H.:Real time robust l1tracker usingaccelerated proximal gradient approach.In:2012IEEE Conference on IEEE Computer Vision and Pattern Recognition(CVPR),pp.1830–1837(2012)4.Candes,E.J.,Tao,T.:Decoding by linear programming.IEEETrans.Inf.Theory51(12),4203–4215(2005)5.Chen M.,L.Z.,Y.,M.:The augmented lagrange multipliermethod for exact recovery of corrupted low-rank matrices.In: arXiv:1009.5055(2010)6.Donoho,D.:Compressed sensing.IEEE Trans.Inf.Theory52(4),1289–1306(2006).doi:10.1109/TIT.2006.8715827.Gu,J.,Nayar,S.,Grinspun,E.,Belhumeur,P.,Ramamoorthi,R.:Compressive structured light for recovering inhomogeneous par-ticipating media.IEEE Trans.Pattern Anal.Mach.Intell.99,1 (2013).doi:10.1109/TPAMI.2012.1308.Hager,G.,Belhumeur,P.:Efficient region tracking with parametricmodels of geometry and illumination.IEEE Trans.Pattern Anal.Mach.Intell.20(10),1025–1039(1998).doi:10.1109/34.722606123Improved 1-tracker using robust PCA...9.Huang,J.,Huang,X.,Metaxas,D.:Learning with dynamic groupsparsity.In:2009IEEE12th International Conference on Computer Vision,pp.64–71(2009).doi:10.1109/ICCV.2009.545920210./PETS2001/:11.Kalal,Z.,Matas,J.,Mikolajczyk,K.:Online learning of robustobject detectors during unstable tracking.In:2009IEEE12th International Conference on Computer Vision Workshops(ICCV Workshops),pp.1417–1424(2009).doi:10.1109/ICCVW.2009.545744612.Mei,X.,Ling,H.:Robust visual tracking and vehicle classificationvia sparse representation.IEEE Trans.Pattern Anal.Mach.Intell.33(11),2259–2272(2011).doi:10.1109/TPAMI.2011.6613.Ng,W.,Li,J.,Godsill,S.,Vermaak,J.:Tracking variable numberof targets using sequential monte carlo methods.In:2005IEEE/SP 13th Workshop on Statistical Signal Processing,pp.1286–1291 (2005).doi:10.1109/SSP.2005.162879414.Ross,D.A.,Lim,J.,Lin,R.S.,Yang,M.H.:Incremental learningfor robust visual tracking(2008)15.Wang,D.,Lu,H.,Yang,M.H.:Online object tracking with sparseprototypes.IEEE Trans.Image Process.22(1),314–325(2013) 16.Wright,J.,Ganesh,A.,Rao,S.,Peng,Y.,Ma,Y.:Robust prin-cipal component analysis:Exact recovery of corrupted low-rank matrices via convex optimization.In:Y.Bengio,D.Schuurmans, fferty,C.Williams,A.Culotta(eds.)Advances in Neural Infor-mation Processing Systems22,pp.2080–2088.Curran Associates, Inc.(2009)17.Wright,J.,Yang,A.,Ganesh,A.,Sastry,S.,Ma,Y.:Robust facerecognition via sparse representation.IEEE Trans.Pattern Anal.Mach.Intell.31(2),210–227(2009).doi:10.1109/TPAMI.2008.7918.Yu,Q.,Dinh,T.B.,Medioni,G.:Online tracking and reacquisitionusing co-trained generative and discriminative trackers.In:Pro-ceedings of the10th European Conference on Computer Vision: Part II,ECCV’08,pp.678–691.Springer-Verlag,Berlin,Heidel-berg(2008).doi:10.1007/978-3-540-88688-4_5019.Zhang,K.,Zhang,L.,Yang,M.H.:Real-time compressive tracking.In:Computer Vision-ECCV2012,pp.864–877.Springer(2012) 20.Zhang,T.,Ghanem,B.,Liu,S.,Ahuja,N.:Robust visual track-ing via structured multi-task sparse put.Vis.101(2),367–383(2013)Dongjing Shan received his B.E.degree in2010,and M.E. degree in Information Science and Technology from Peking University,China,in2013.He joined the Chongqing Key Lab-oratory of Emergency Commu-nication in2013and is now a researcher in that Laboratory.His research interests include com-puter vision,image processing and patternrecognition.Zhang Chao received his Ph.D. degree in Electrical Engineering in1995,from Northern Jiao-tong University,Beijing,China. After working as a postdoctoral research fellow for two years, he became a faculty member in June1997at the State Key Laboratory of Machine Percep-tion,Peking University,Beijing, China,where he is currently an associate professor of machine vision research group.123。