Review

Human physiome based on the high-resolution datasetof human body structure

Qian Liu, Bo Wu, Shaoqun Zeng, Qingming Luo *

Britton Chance Center for Biomedical Photonics, Wuhan National Laboratory for Optoelectronics,

Huazhong University of Science and Technology, Wuhan 430074, China

Received 13 December 2007; received in revised form 23 March 2008; accepted 23 March 2008

Abstract

Physiome Project, as a new concept, is proceeding rapidly with the great advancement of genomics, physiological experiment, biologicmodeling and computer simulation technique. The project seeks to provide a quantitative framework for modeling of the human physio-logical system using computational approaches, which is able to integrate the knowledge of molecular biology, biochemical, biophysicaland anatomical information on different levels, including cell, tissue, organ, system and organism. This paper reviews the development ofthe Physiome Project in the past decade. The role of high-resolution datasets of human body structure in Physiome Project is discussed. Thefuture plan and applications of the high-resolution datasets of human body structure to Physiome Project are discussed as well.� 2008 National Natural Science Foundation of China and Chinese Academy of Sciences. Published by Elsevier Limited and Science inChina Press. All rights reserved.

Keywords: Physiome; Physiome Project; Multi-scale modeling; Anatomical modeling

1. Introduction

The development of biology has been greatly acceler-ated by advanced biotechnology and the extensive useof computers in recent decades. Integration and mathe-matical modeling of the extremely large datasets ofsequence, protein structure and signal-transduction-path-way, which was generated from Genome [1–3] and Pro-teome research, is necessary to further study thephysiological function of complicated cells, tissues andorgans [4]. Independently, progress in the Visible HumanProject (VHP) in the USA, China and Korea [5] hasoffered innovative approaches, including cyber-basedcomputing, visualization and simulation, to representthe realistic human anatomy [6], where the structuraldata have been obtained, digitized, reconstructed andvisualized for individual tissues, organs and systems,

respectively [7]. A comprehensive and informative frame-work, scaling from structure to function and from geneto the whole organism, is therefore urgently needed tointegrate these two approaches [8].

Integration of the molecular biology, biochemical, bio-physical and anatomical information of structure andfunction at different levels has been enabled in ‘‘Physiom-e” by a wide use of computational biology method andcomputer technique. ‘‘Physiome” was presented in areport from the Commission on Bioengineering in Physi-ology to the 32nd World Congress Council of Interna-tional Union of Physiological Sciences in Glasgow, UK,in 1993, then Jim Bassingthwaighte introduced the concep-tion of ‘‘Physiome” in 1995 [9]. The term ‘‘Physiome”

originates from ‘physio’ (denote physiology) and ‘ome’(as a whole) [10,11], it will provide a framework for mod-eling the human body and adopt a group of computa-tional models to analyze integrative biological functionand to provide a system for clinical medicine hypothesissimulation and experimentation testing [12,13].

1002-0071/$ - see front matter � 2008 National Natural Science Foundation of China and Chinese Academy of Sciences. Published by Elsevier Limited

and Science in China Press. All rights reserved.

doi:10.1016/j.pnsc.2008.03.013

* Corresponding author. Tel.: +86 27 87792033; fax: +86 27 87792034.E-mail address: qluo@mail.hust.edu.cn (Q. Luo).

www.elsevier.com/locate/pnsc

Available online at www.sciencedirect.com

Progress in Natural Science 18 (2008) 921–925

We have used the datasets and methods of the Phys-iome Project to develop a model integrating structuraland functional information about cells, tissues andorgans based on the anatomical dataset of the VisibleChinese Human Project. This has been applied to nuclearmedicine, biomechanics and biomedical optics relatedfields as described below.

2. The Physiome Project

The Physiome Project is an internationally collaborativeopen-source project to provide a public domain frameworkfor computational physiology, including the developmentof modeling standards, computational tools and web-accessible databases of models of structure and functionat all spatial scales such as cell, tissue and organ level[14]. The goal of the project is to use computational mod-eling to simulate integrative biological function at all thelevels and to use the models to bring benefits to the designof medical devices, and to the diagnosis and treatment ofdisease [15].

International Union of Physiological Sciences (IUPS) isthe primary sponsor and planner of the Physiome Project.There are several key institutions involved in this project,including the University of Auckland in New Zealand,the University of Oxford in the UK, Washington Univer-sity and John Hopkins University in the USA [16]. Bynow this project has achieved some successes, such as link-ing anatomical structure and molecular and cellular eventswith physiological function with multi-length scales andmulti-time scales; developing a group of computer lan-guages based on XML [17] technology to describe biologi-cal structure and function at different levels [18,19]; andbuilding physiological models at organ and tissue levels[20,21].

3. Multi-scale modeling of the human physiology

The Physiome Project plans to develop an infrastruc-ture for linking models of biological structure and func-tion in human across multiple levels of spatialorganization and multiple time scales. The componentsof Physiome multi-level architecture include gene regula-tion network, protein interaction network, metabolicpathway, subcellular signaling cascades, organ and tissuestructure and physiological function. Any attempt to linkmolecular, cellular and physiological events with physio-logical function must deal with length scales that rangefrom the 1 nm that is typical of a protein to the 1 mscale of an intact body. Similarly, the range of timescalesshould encompass the 1 ls that is ion transmembranetransport to the 108 s that is the characteristic of somediseases such as cancer [10]. So a hierarchy of modelsis needed, different types of model are appropriate to dif-ferent levels including molecule, cell, tissue, organ, organsystem and organism. Several standard markup lan-guages such as CellML, TissueML and OrganML were

developed to describe each level’s data and link parame-ters between two levels.

The complexity of biological systems lies in the struc-tural and functional information from system to organand from tissue to cell as well as the relationship betweenthem. Highly integrated and effective managed platformshave been developed to serve the multiple functions ofmodeling, visualization, and further analysis. The CMISS(Continuum Mechanics, Image analysis, Signal processingand System Identification) [22], which is an interactivemathematical modeling environment developed by Hunteret al. (2001) [23], facilitates the application of finite elementanalysis, boundary element and collocation techniques to avariety of complex problems in cardiac electrophysiologysimulation [24], which involves calculations on both cellu-lar and metabolic levels and three-dimensional visualiza-tion of anatomical model at organ level. CMISS treatsthe physiological data in format of FieldML. It comprisestwo major modules, including a computational core namedCM that supports remote control on powerful worksta-tions or supercomputers and a graphical interface capableof advanced 3D displaying named CMGUI, which pro-vides services for 3D visualization and manipulation.CMISS has been widely applied in linear and nonlinearsimulations of mechanics, electrics and hydrodynamics rel-evant to cardiac electrophysiology.

4. Progress in physiome modeling

The organ level modeling is the central problem ofPhysiome Project, which is pivot in the multi-scale mod-eling and has been deeply studied. An early and excellentwork at organ level is the integrated model of the heart,which was developed by collaborative work under theleadership of Hunter at Auckland University. The heartof a dog specimen was sectioned into thousands of extre-mely thin slices. A three-dimensional model was able tobe reconstructed based on the cryosectional information[25]. A computational framework was released for inte-grating the electrical, mechanical and biochemical func-tions of the heart. Finite element techniques were usedto solve the large-deformation soft tissue mechanics usingorthotropic constitutive laws based on the measuredfibre-sheet structure of myocardial. The reaction–diffu-sion equations governing electrical current flow in theheart were solved on a grid of deforming material pointswhich represent the cellular processes underlying the car-diac action potential [23]. In 2002, Professor Noble’sgroup in Oxford University proposed the concept of con-tinuity modeling the heart from gene to organ level [20].They tried to interpret the relationship between geneexpression and organ physiological function. For thispurpose, they developed an electrophysiological continu-ity model that consists of multiple processes from geneto cardiac events to explain why the expression of someabnormal genes can lead to heart diseases such asarrhythmia [26]. A Chinese research group led by Profes-

922 Q. Liu et al. / Progress in Natural Science 18 (2008) 921–925

sor Wang in Xiamen University developed a model forhuman eyes. They obtained image data which contain646 slice images of eye in February 2004 and developeda new approach that is based on slit-lamp imaging forthe surface reconstruction of corneal [27]. The futureplan is to build a ‘‘virtual-eye” at tissue, cell, even mol-ecule level and use this model in analysis of physiologicalfunction process and drug design.

5. The Visible Chinese Human Project

Modeling-on-structure is an important specialty ofPhysiome. The Visible Human Project (VHP) is a greatachievement in modeling the human body anatomicstructure, which was initiated by US National Libraryof Medicine in 1989. Health Science Center of the Uni-versity of Colorado collected a structural dataset ofhuman body after successive sectioning and photograph-ing. A complete digital image dataset of a Caucasianadult male was obtained and made available on-line forpublic use in November 1994 [28]. Ever since then, num-bers of researches on digital human have emerged ininstitutions all over the world for the significant impactand widespread application of VHP. The Visible ChineseHuman (VCH) Project was initiated for the purpose ofestablishing a virtual Chinese human model. SouthernMedical University carried out the specimen selectionand sectioning work, and the resulting datasets were thensent to Huazhong University of Science and Technologyfor image processing and visualization. This work wascompleted in 2005. The high-quality image set was seg-mented, registered, and identified interactively. The wholebody was reconstructed and visualized by parallel com-puting on a cluster. The anatomic structures at tissue,organ and organ system levels were labeled and anno-tated in category. A high-quality 3D human organic sys-tem database was constructed [29–32]. Fig. 1 shows theanatomical structure for a whole human body. Com-pared with the VHP and Visible Korean Human Project[33], our digital color images of every successive slice at5440 � 4080 pixel resolution correspond to a voxel sizeof 0.1 � 0.1 � 0.2 mm, which is the highest resolutionat present [34].

6. Multi-scale modeling of high-resolution structure dataset

Multi-scale modeling is a characteristic in physiomeresearch. We developed a quantitative framework toachieve multi-scale modeling on the structure and functionof cells, tissues, and whole organs.

Fig. 2 shows the framework of multi-scale modeling. Atthe organ and organ system level, 3D anatomical modelsof nine physiological systems have been constructed, includ-ing respiratory system, reproductive system, endocrinesystem, cardiovascular and musculoskeletal system, circula-tory system, digestive system, nervous system, sensory sys-tem, urinary system, and 140 subsystems. A group of

radiological dosimetry models [35,36], musculoskeletal sys-tem biomechanics models, and tissue optical models werealso constructed based on the structure dataset. PhysioMLand AnatML markup languages were employed to describeand store the data. At the tissue and subcellular levels, theOCT, confocal microscopy, and ultramicrotome microscopysystem were adopted to get structure information of organ-ism tissues [37,38], the data were classified into four typeswhich are muscle tissue, connective tissue, nerve tissue, andepithelial tissue. CellML and TissueML markup languageswere utilized to describe name, type, chemical reaction andphysics parameters of each level’s data from cell to tissue.At the protein and molecule levels, sequence and domaininformation of gene and protein was obtained from Gene-Bank, EMBL, and DDBJ, these data were treated in formatof FASTE [39]. A platform – Biolab [40] which is based oncomputational biology can incorporate the information ofeach level. A markup language was developed to relateparameters at different levels [41]. The physical and structuremodels (VRML format) were imported into a MeSH ontol-ogy database and supply service to platform user [42]. Avisual online resource was also provided for biologists tomodel, simulate, and compute from the web [32]. The pri-mary aim of this Physiome research platform is to achievethe incorporation, computing and simulation of structure,

Fig. 1. Three-dimensional anatomical modeling for a human body.

Q. Liu et al. / Progress in Natural Science 18 (2008) 921–925 923

and functional information of biological system at differentlevels.

7. Physiome modeling based on high-resolution structure

dataset

Based on the high-resolution dataset of VCH, we con-ducted a study on the Physiome, including physiologicalmodeling and simulation in nuclear medicine, biomechan-ics and biomedical optics.

Accurate determination of the radiation exposure hasalways been an important task in the field of radiolog-ical sciences. The VCH computational model has beensuccessfully implemented to Monte Carlo dosimetry cal-culations for external photons. Irradiation sources weredefined in six standard positions to calculate the equiv-alent radiate translate modulus and absorbed doses fordifferent radiate energy. Measurements of radiate dosesof 48 organs and tissues of whole human body of pho-ton, neutron and proton have completed [35,36]. Thiswork provides a high-resolution data for nuclear radio-protection simulation. In the field of biomechanics, wereconstructed the voxel-based model, divided voxelmodel to finite element model, and gave its organiza-tional material properties. Various kinds of mechanicalparameters such as anisotropy, elasticity modulus, Pois-son’s ratio, material consistency, contact quality, andelastic limit of skeletal system have been evaluated.Other organ’s mechanical characteristics study is stillunderway. The above results can be used in theresearch of exercise physiology, for example, evaluatethe human body damnification in accident [37] andsimulate corrective action of athlete. We plan to use3D photon migration algorithm to simulate processes

of photon creating, moving, dispersing, absorbing anddemising in human tissue [38] based on the VCHdataset.

8. Prospect

The physiome framework which we show in this reviewis huge. It ranges from modeling molecular events to mod-eling whole organism. The ‘‘Multi-scale modeling” makesit possible to incorporate vast amount of information atthe level of genes, proteins, cells, tissues and organs. How-ever, it is necessary to emphasize the enormity of the chal-lenge in further study. First, more precise and morequantitative mathematical models are required to interpretthe relationship between structure and function at differentlevels, which need the support of mass of experimental dataand more detailed biological knowledge. Second, the devel-opment of database standards which allow the easyexchange and share of data between different researchgroups is an important task for international communion.Third, Physiome markup languages which based on XMLmust be developed to encapsulate models at all levels.

The high-resolution dataset of a human body structureis a basis of Physiome research, which has many applica-tions in the field of medical diagnostics, virtual surgery,education and medical training [43]. The three-dimensionalsurgical simulator that is designed based on the body struc-ture dataset can teach surgeons the visuospatial skillsrequired to navigate through a transpetrosal approach[44]. The process of minimally invasive surgery can be sim-ulated and the effect can be predicted and analyzed. There-fore, this work could improve the security and success rateof the surgery [45]. The radiological dosimetry models, bio-mechanics models, and tissue optical models, which were

Fig. 2. Multi-scale modeling of the human physiology.

924 Q. Liu et al. / Progress in Natural Science 18 (2008) 921–925

constructed based on the structure datasets, can be directlyused in the field of nuclear medicine, exercise physiology,and biomedical optics to promote the development of thesesubjects.

Acknowledgements

This work was supported by the National High-TechResearch and Development Program of China (2006AA02Z343). The authors acknowledge members of the ‘‘VisibleHuman” research group in Britton Chance Center for Bio-medical Photonics, Huazhong University of Science andTechnology, in Wuhan, China, for their excellent work.

References

[1] Venter JC, Adams MD, Myers EW, et al. The sequence of the humangenome. Science 2001;291(5507):1304–51.

[2] McPherson JD, Marra M, Hillier L, et al. A physical map of thehuman genome. Nature 2001;409(6822):934–41.

[3] Collins FS, Green ED, Guttmacher AE, et al. A vision for the futureof genomics research. Nature 2003;422(6934):835–47.

[4] Hunter PJ, Borg TK. Integration from proteins to organs: thePhysiome Project. Nat Rev Mol Cell Biol 2003;4(3):237–43.

[5] Federation of American Scientists (FAS)’s Digital Human Project.Available from: http://www.fas.org/main/content.jsp?formAction=325&projectID=12 [2007-8-12].

[6] Liu Q, Gong H, Luo QM. Parallel visualization of visible Chinesehuman with extremely large datasets. In: 27th IEEE EMBS AnnualInternational Conference, vol. 5, Shanghai; 2005. p. 5172–5.

[7] Gagvani N, Silver D. Animating volumetric models. Graph Models2001;63(6):443–58.

[8] Noble D. The rise of computational biology. Nat Rev Mol Cell Biol2002;3(6):460–3.

[9] Bassingthwaighte JB. Toward modeling the human physionome. AdvExp Med Biol 1995;382:331–9.

[10] Hunter PJ, Borg TK. Integration from proteins to organs: thePhysiome Project. Nat Rev Mol Cell Biol 2003;4(3):237–43.

[11] Hunter PJ. The IUPS Physiome Project: a framework for computa-tional physiology. Prog Biophys Mol Biol 2004;85(2–3):551–69.

[12] Hunter PJ, Robbins P, Noble D. The IUPS human physiome project.Eur J Physiol 2002;445(1):1–9.

[13] Crampin EJ, Halstead M, Hunter PJ, et al. Computational physiol-ogy and the physiome project. Exp Physiol 2003;1(89):1–26.

[14] Bassingthwaighte JB. The macro-ethics of genomics to health: thephysiome project. CR Biol 2003;326(11):1105–10.

[15] Bassingthwaighte JB. Strategies for the physiome project. AnnBiomed Eng 2000;28(8):1043–58.

[16] IUPS Physiome Project Roadmap. Available from: http://www.phys-iome.org.nz/roadmap/roadmap-mar05 [2007-8-12].

[17] Achard F, Vaysseix G, Barillot E. XML, bioinformatics and dataintegration. Bioinformatics 2001;17(2):115–25.

[18] Lolyd CM, Halstead MD, Nielsen PF. CellML: its future, presentand past. Prog Biophys Mol Biol 2004;85(3):433–50.

[19] Hucka M, Finney A, Sauro HM, et al. The systems biology markuplanguage (SBML): a medium for representation and exchange ofbiochemical network models. Bioinformatics 2003;19(4):524–31.

[20] Noble D. Modeling the heart – from genes to cells to the whole organ.Science 2002;295(5560):1678–82.

[21] Hunter PJ, McCulloch AD, Keurs HE. Modeling the mechanicalproperties of cardiac muscle. Prog Biophys Mol Biol 1998;69:289–331.

[22] An interactive computer program for continuum mechanics, imageanalysis, signal processing and system identification. Available from:http://www.cmiss.org [2007-8-12].

[23] Christie GR, Bullivant DP, Blackett SA, et al. Modelling andvisualising the heart. Comput Vis Sci 2002;4(4):227–35.

[24] Smith NP, Nickerson DP, Crampin EJ, et al. Multiscale computa-tional modelling of the heart. Acta Numer 2004;13:371–431.

[25] Tomlinson KA. Finite element solution of an eikonal equationfor excitation wavefront propagation in ventricular myocar-dium. Ph.D. Thesis, The University of Auckland, NewZealand; 2000.

[26] Noble D. Unraveling the genetics and mechanisms of cardiacarrhythmia. Proc Natl Acad Sci USA 2002;99(9):5755–6.

[27] Zhuge B, Zhao C, Wang BL, et al. A new approach to the surfacereconstruction of corneal based on slit-lamp imaging. Chin J BiomedEng 2005;24(3):353–6.

[28] Visible Human Project. Available from: http://www.uchsc.edu/sm/chs/open.html [2007-8-12].

[29] Tang L, Li ZY, Liu C, et al. Establish high resolution imagedataset of Chinese digital human male. J Med Biomech2006;21(3):179–82.

[30] Li AA, Liu Q, Gong H, et al. High quality 3D skeleton systemmodeling of virtual Chinese human male No.1. Chin J Clin Anat2006;24(3):292–4.

[31] Wang WJ, Liu Q, Gong H, et al. 3D Reconstruction of cardiovas-cular system on virtual Chinese male-No.1. J Med Biomech2006;21(3):198–202.

[32] Visible Chinese Human. Available from: http://www.vch.org.cn[2007-8-12].

[33] Visible Korean Human. Available from: http://vkh3.kisti.re.kr [2007-8-12].

[34] China National Center for Biotechnology Development. Availablefrom: http://www.cncbd.org.cn/gndt/gndt/060208/01.html [2007-8-12].

[35] Zhang G, Luo QM, Zeng SQ, et al. The development and applicationof the visible Chinese human model for Monte Carlo dose calcula-tions. Health Phys 2008;94(2):118–25.

[36] Zhang G, Liu Q, Luo Q. Monte Carlo simulations for externalneutron dosimetry based on the visible Chinese human phantom.Phys Med Biol 2007;52:7367–83.

[37] Lu Q, Zeng S, Luo QM, et al. Rapid modeling of two-dimensionalmulti-photon excitation microscopic imaging through turbid med-ium. Opt Commun 2001;189(4–6):227–34.

[38] Zhou SW, Zhang ZH, Liu Q, et al. The application of opticalimaging: studying the photodynamic effect of tumor in vivo. ProgBiochem Biophys 2005;32(10):969–74.

[39] Wang P, Lu Q, Luo QM, et al. Building bioinformatics secondarydatabase with web information automatic retrieval. Chin HighTechnol Lett 2004;14(5):28–31.

[40] Luo QM. A plantform for bioinformatics simulate and visualization.China Patent 2005SR01010, 2004.

[41] Dong JQ, Lu Q, Luo QM, et al. Java and XML based graphicsoftware package for modeling biochemical and gene regulatorynetworks. Comput Eng Appl 2004;40(6):202–4.

[42] Han D, Shu XQ, Zeng SQ, et al. Interactive 3D atlas system ofhuman anatomy based on database. Chin J Med Phys2008;25(1):492–5.

[43] Zhong SZ. Digital human and digital anatomy. Jinan: ShandongScience and Technology Press; 2004, [p. 613–5].

[44] Bernardo A, Preul MC, Zabramski JM, et al. A three-dimensionalinterative virtual dissection model to simulate transpatrous surgicalavenues. Neurosurgery 2003;52(3):499.

[45] Zhong SZ. Actualities and prospects of research on digitizedvirtual human. Med J Chin People’s Liberation Army2003;28(5):383–5.

Q. Liu et al. / Progress in Natural Science 18 (2008) 921–925 925