SPECIAL ARTICLE www.jasn.org

GUDMAP: The Genitourinary Developmental Molecular Project

Andrew P. McMahon,* Bruce J. Aronow,† Duncan R. Davidson,‡ Jamie A. Davies,§ ʈ Kevin W. Gaido, Sean Grimmond,¶ James L. Lessard,** Melissa H. Little,¶ S. Steven Potter,** Elizabeth L. Wilder,†† and Pumin Zhang,‡‡ for the GUDMAP project

*Department of Molecular and Cellular Biology and Harvard Stem Cell Institute, Harvard University, Cambridge, Massachusetts; †Division of Biomedical Informatics, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio; ‡Medical Research Council Human Genetics Unit, Western ʈ General Hospital, and §Centre for Integrative Physiology, University of Edinburgh, Edinburgh, United Kingdom; Hamner Institutes for Health Sciences, Research Triangle Park, North Carolina; ¶Institute for Molecular Bioscience, University of Queensland, St. Lucia, Queensland, Australia; **Division of Developmental Biology, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio; ††Renal and Urogenital Development, Kidney Injury and Repair, and Basic Science of Cystic Kidney Disease Research, National Institutes of Health Roadmap Interdisciplinary Research Working Group, Bethesda, Maryland; and ‡‡Department of Molecular Physiology and Biophysics, Baylor College of Medicine, Houston, Texas

ABSTRACT toward the treatment of disease have In late 2004, an International Consortium of research groups were charged with the task thrown a spotlight on the normal develop- of producing a high-quality molecular anatomy of the developing mammalian urogenital mental programs that orchestrate develop- tract (UGT). Given the importance of these systems for human health and repro- ment of our organ systems. Which cells duction, the need for a systematic molecular and cellular description of their developmen- make up an organ? How are they gener- tal programs was deemed a high priority. The information obtained through this initiative ated? Which factors control these events in is anticipated to enable the highest level of basic and clinical research grounded on a time and space, ensuring a co-herent, func- 21st-century view of the developing anatomy. There are three components to the Geni- tional system? How are organs repaired? tourinary Developmental Molecular Anatomy Project GUDMAP; all of these are intended What controls the numbers of any given to provide resources that support research on the kidney and UGT. The first provides cell type and the overall balance of cell ontology of the cell types during UGT development and the molecular hallmarks of those number in the organ? Addressing any of cells as discerned by a variety of procedures, including in situ hybridization, transcriptional these pertinent and fascinating questions profiling, and immunostaining. The second generates novel mouse strains. In these strains, requires a solid foundation in develop- cell types of particular interest within an organ are labeled through the introduction of a mental anatomy. The overarching goal of specific marker into the context of a gene that exhibits appropriate cell type or structure- the Genitourinary Developmental Molec- specific expression. In addition, the targeting construct enables genetic manipulation ular Anatomy Project (GUDMAP) initia- within the cell of interest in many of the strains. Finally, the information is annotated, tive is to establish this foundation for the collated, and promptly released at regular intervals, before publication, through a data- UGT. The ensuing data generated by the base that is accessed through a Web portal. Presented here is a brief overview of the initiative will enable, facilitate, and stimu- Genitourinary Developmental Molecular Anatomy Project effort. late research into, and our understanding of, this critical organ network. J Am Soc Nephrol 19: 667–671, 2008. doi: 10.1681/ASN.2007101078

One in nine Americans has chronic kid- velopmental deficiencies likely underlie or Published online ahead of print. Publication date ney disease (CKD), and the same num- influence susceptibility to a range of dis- available at www.jasn.org. ber is at increased risk for developing eases of the urogenital tract (UGT), as well Correspondence: Dr. Andrew P. McMahon, Depart- CKD. Approximately 12% of women in as adversely affecting fertility. One example ment of Molecular and Cellular Biology and Harvard the United States have a problem conceiv- is the link between nephron number, de- Stem Cell Institute, Harvard University, 16, Divinity Ave- ing or bringing a baby to term. Approxi- termined at the outset by a fetal develop- nue, Cambridge, MA 02138. Phone: 617-496-3757; Fax: 617-496-3763; E-mail: [email protected]; mately 3% of all children are born with a mental program, and hypertension and re- consortium Web address: http://www.gudmap.org recognized birth defect. Some of the most nal disease later in life.1 E.L.W.’s current affiliation is Office of Portfolio Anal- common are those that affect the genito- This increased realization of a link be- ysis and Strategic Initiatives, National Institutes of urinary system, with cryptorchidism and tween development and disease and the Health, Bethesda, Maryland. hypospadias being the most common male opportunities that might come from ex- Copyright © 2008 by the American Society of birth defect. Furthermore, less obvious de- ploiting our developmental understanding Nephrology

J Am Soc Nephrol 19: 667–671, 2008 ISSN : 1046-6673/1904-667 667 SPECIAL ARTICLE www.jasn.org

INTRODUCTION TO GUDMAP by the NIDDK and the National Institute of approach to map gene expression do- Child Health and Human Development. mains and identify patterns that are diag- In September 2001, the National Insti- nostic of specific anatomic regions or in- tute of Diabetes and Digestive and Kid- dividual cell types. WISH has been used ney Diseases (NIDDK) was advised by a AN ANATOMIC ONTOLOGY FOR generally to survey expression of specific working group of its National Advisory THE ANNOTATION OF THE classes of genes within the UGT. Tran- Council to form a strategic plan to con- DEVELOPING UGT scriptional regulators, as a functional sider how stem cells and developmental category, are valuable cell type–specific biology might be used to explore the re- Early discussions among the members of markers in several developmental stud- pair and replacement of damaged organs the consortium highlighted the need to ies.3 Furthermore, their known regula- and to determine how they might pro- develop a high-resolution anatomy- tory actions can provide insights into vide insights into the pathologic pro- based ontology of the UGT. This ontol- possible actions within expressing cells, cesses underlying developmental defects ogy would not only provide a common stimulating further experimental in- and disease. A separate working group of descriptive language for cell types within quiry. Optimizing a WISH procedure for the National Advisory Council high- the developing UGT, an essential prereq- sensitivity, penetration, and specificity, a lighted developmental biology as an un- uisite for any search algorithm and the systematic, genome-scale screen was per- derstudied area of bladder research. logical construction of the critical GUD- formed for the spatial expression of a These reports led the NIDDK to convene MAP database, but also draw these terms majority of mouse transcriptional fac- a panel of advisors in January 2003 to into a developmental series of hierarchi- tors in at embryonic day 15.5 UGT (Yu et discuss approaches that would provide a cal relationships among the cell popula- al. in preparation; http://www.gudma- fundamental description of the develop- tions of the UGT. M.H.L. spearheaded p.org/Research/Protocols/). The anno- ing UGT. the important goal of a high-resolution tated data representing approximately The working group recommended anatomic ontology of the UGT in con- 1500 mouse transcriptional factors can several strategies to reach this goal: (1)A sultation with GUDMAP members, be queried through the GUDMAP data- high-throughput in situ analysis of gene members of the EuReGene network base. In addition, an analysis of the ex- expression in the developing UGT, (2)a (http://www.euregene.org), and other pression pattern of genes classified as ex- high-resolution analysis of a limited set international experts in the field. The on- tracellular on the basis of gene ontology of genes in time and space to define the tology was published in 2007.2 This on- terms and previously described to be ex- emergence of anatomic and functional tology provides a framework for anno- pressed during kidney development has domains, and (3) the creation of a data- tating the expression data. Furthermore, been submitted.4 A second comprehen- base to make these data readily and the expression studies provide an ongo- sive genome-scale screen of mammalian promptly available to the research com- ing test and challenge to the ontology. signals and their receptors is under way, munity. The ontology will inevitably evolve as together with prioritized gene sets based Requests for applications were issued molecular approaches reveal molecularly on spatial expression profiling of the de- in late 2003, and a funded consortium of distinct cell types that are not currently veloping UGT (see below). All data are seven laboratories was formed from the incorporated into -based anat- released on a regular basis through the applicants late in 2004. The group con- omy. We urge that this ontology be GUDMAP database. Together, these sisted of six data-generating centers: Five adopted broadly within the research screens have identified a large number of in the United States (A.P.M., K.W.G., community. A common standard for an- genes that have spatially discrete patterns J.L.L., S.S.P., and P.Z.), one in Australia notating data will maximize the value of of expression and map to distinct re- (S.G. and M.H.L.) and a database in the that data. For example, there will be a gional domains within each organ sys- United Kingdom (D.R.D. and J.A.D.). better understanding as to which cell tem. Subsequently, consideration of the Stra- type is being discussed if the terms are Furthermore, genes can be clustered tegic Plan for Pediatric Urology, devel- defined and accepted. Furthermore, this into groups on the basis of these patterns. oped by the NIDDK in 2006, and addi- would more readily facilitate the incor- The three-dimensional view afforded by tional advice from an external panel of poration of data generated outside of the WISH analysis is particularly valuable in expert consultants led the NIDDK and GUDMAP consortium at some time in identifying specific patterns that might the National Institute of Child Health the future. be harder to identify in section analysis. and Human Development to add three The low-resolution WISH data also pro- collaborators to the group to maximize the vide an excellent prescreen for further effectiveness and coverage of the GUD- MOLECULAR ANATOMY OF THE high-resolution section in situ hybridiza- MAP effort: Two in the United States DEVELOPING UGT tion (SISH) analysis, increasing the like- (B.J.A. and Michelle Southard-Smith) and lihood that any gene will give useful in- the other in Australia (Peter Koopman). Whole-mount in situ hybridization formation. GUDMAP is a 5-yr project jointly funded (WISH) is being used as a low-resolution SISH and immunohistochemistry

668 Journal of the American Society of Nephrology J Am Soc Nephrol 19: 667–671, 2008 www.jasn.org SPECIAL ARTICLE

(IHC) provide high-resolution ap- analyzed using these approaches. The ontology to include domains that are proaches that enable gene activity to be primary transcriptional profiling data identifiable only via gene expression. mapped at single-cell or near-single-cell and the emerging expression studies Figure 1 provides an overview of the resolution within a (http://ww- based on these data are available on the workflow and data generation within w.gudmap.org/Research/Protocols/). GUDMAP Web site. Profiling is also be- GUDMAP. Although individual groups SISH also identifies synexpression clus- ing applied to mutant mouse strains in may have a more organ-specific interest, ters—genes that are “coexpressors”— which the primary developmental defect the consortium has attempted to maxi- prioritizing these for in-depth annota- is known but the underlying mechanism mize the effectiveness of the entire tion of specific domains. Large-scale remains to be fully elucidated. For exam- group’s effort to define UGT develop- SISH screens have formed the basis for ple, Wnt4 mutants fail to generate epi- ment molecularly. molecular of the developing thelial renal vesicles, the precursor for mouse nervous system3 and the adult the main body of the nephron.12 Thus, brain,5 enhancing research on the central Wnt4 mutants form the basis of a screen MOUSE STRAINS FOR THE nervous system. High-throughput SISH that enriches for genes representative of ANALYSIS OF UGT studies are under way within several the renal vesicle or its more differenti- DEVELOPMENT groups within GUDMAP addressing ated nephron derivatives. Identifying each component part of the UGT. A genes whose expression is downregu- Transgenic approaches enable the gener- complementary, comprehensive initia- lated in Wnt4 mutants identified several ation of novel mouse strains in which key tive for the kidney has been initiated hundred candidates. Analysis of their ex- structures are labeled in an organ of in- within the EuReGene consortium pression showed that half are restricted terest.14 The generation of novel mouse (http://www.euregene.org). to Wnt4-dependent structures.13 Thus, strains is an important part of the GUD- The analysis of bulk transcriptional the screen significantly enriches for genes MAP initiative. The consortium goal has activity through microarray-based tech- that demarcate the target component of been to identify cell-type or anatomic re- nology is providing unprecedented in- the kidney. In both instances, accompa- gions of interest within the UGT by the formation into cell-, tissue-, and organ- nying WISH and SISH of the resultant expression of a specific gene. Gene tar- type gene expression programs. gene lists acts both to validate the data geting is then used to introduce, under Transcriptional analysis of the wild-type and to identify novel subcompartments the control of the cell/tissue-type marker UGT at different stages of develop- of the structures initially profiled, gene, expression constructs that enable ment4,6–8 and within specific cellular or thereby expanding our developmental these cells to be identified, isolated, and structural subcompartments of those or- gans4,8–10 has provided temporal and spatial profiles of gene expression where de novo expression of a gene may poten- tially correlate with the emergence of new cell identities and distinct struc- tures. In GUDMAP, microarray profil- ing has been used in several ways. In the first instance, increased spatial resolution is being sought through the direct isolation of specific cell types or anatomic structures and their subse- quent transcriptional profiling. This provides detailed insight into local gene expression programs. Cell populations of interest have been isolated by manual dissection, by laser capture microsco- Figure 1. Data flow in the GUDMAP. Gene expression results from microarray and in situ py,11 or through FACS where markers assays are submitted by consortium laboratories to a central editorial office. Microarray are available (see below). After amplifi- data includes .cel, .chp, .rpt, and .txt files. In situ data include images, definitive descrip- cation of resulting cDNA, the gene ex- tions of probes, and experimental details; gene expression patterns are annotated with a standard anatomic ontology by the data-producing laboratories. In situ and microarray pression profile is determined predomi- submissions are curated by editors and entered into the database within a few days of nantly through the use of high-density receipt. For rapid data release, uncurated microarray files are made available for down- Affymetrix arrays. load from an archive Web page on the day of submission. A suite of interfaces to the Currently, profiling data exist for GUDMAP database make the data available to researchers. In addition, other bioinfor- most organs of the UGT. Within the kid- matics resources will have direct programmatic access to the GUDMAP database. Curated ney, 12 cellular compartments have been information about the consortium mouse strains is also available from the Web site.

J Am Soc Nephrol 19: 667–671, 2008 Genitourinary Developmental Molecular Anatomy Project 669 SPECIAL ARTICLE www.jasn.org genetically manipulated. In one example, types for which current tools were inad- tools, and community resources relevant a GFP:Cre fusion protein encoding equate. to the UGT. gene15 is introduced through embryonic Details of the selected genes, targeting The Web-based GUDMAP database stem cell–mediated gene targeting. In events, and progress can be viewed at the has been publicly accessible since April mice resulting from germline transmis- GUDMAP Web site (http://www.gud- 2006. Consortium members submit data sion of the targeted embryonic stem cell map.org/Resources/MouseStrains/). All on a monthly basis to the GUDMAP ed- genome, the cell type of interest can be strains will be made publicly available itorial office, where they are curated and readily identified by GFP fluorescence in through a third-party distribution cen- made available within the database, nor- the native tissue. In addition, cells can be ter. Selection of the center is under way. mally within a few days of receipt. Im- FACS sorted and subjected to high-reso- Community input is welcomed in the se- portantly, the database provides an effec- lution transcriptional profiling, or the lection of cell types for marking, candi- tive way to query and view data fate of the cell type of interest determined date genes for targeting, and vector selec- generated by the consortium. The initial whereby CRE-mediated recombination tion for the generation of new mouse priority was to provide immediate access is used to activate any of a number of Cre strains (http://www.gudmap.org/Con- to data from a variety of simple inter- reporter mouse strains. tact/Mice.html). faces. Accordingly, typing any text term The consortium goal is to generate 30 in the quick search slot enables the que- to 40 new mouse lines. Each of these rying of the entire database. Alterna- strains will undergo an initial analysis tively, entries in the database can be within the consortium to verify that the GUDMAP DATABASE browsed in table form; columns in the reporter expression matches the expres- table can be re-ordered to segregate sion of the endogenous gene to which it The Web-based GUDMAP database (ac- WISH, SISH, IHC, and array data or to is targeted and to confirm the activity of cessible through http://www.gudmap.org) sort data by developmental age/stage, by any additional regulatory components houses all of the data generated by the con- gene symbol, and so forth. Standard (e.g., Cre-based DNA excision). The sortium (Figure 2). This includes primary query interfaces allow researchers to list choice of gene/cell type to be targeted to WISH, SISH, IHC, microarray-based tran- genes expressed in any particular ana- date has been directed either by an exist- scriptional profiling, mouse strain charac- tomic structure or to find the expression ing objective to characterize a specific terization data together with subsequent pattern of any particular gene; an ad- cell population or on the advice of the follow-up analyses of these data sets, and vanced interface lets users build custom research community with respect to cells information regarding methods, research queries to access every field in the data- base. Database users can choose to focus on particular parts of the UGT, for exam- A ple metanephros or female reproductive system. Using any of these functions, re- searchers can then make collections of database entries and explore their inter- section with the results of further que- ries. In this way, for example, one can find genes expressed in the nephrogenic zone in the newborn kidney but not in the nephrogenic zone at embryonic day B Six2 Sox18 15.5. The database is an evolving resource. We anticipate that the database will pro- vide sophisticated views of data to enable researchers to build new hypotheses Sox9 Foxd1 about gene function in the UGT. For ex- ample, a researcher will be able enter a list of genes, perhaps from his or her own work, and view all related data in the da- tabase, enabling him or her to explore pathways and gene interactions on the Figure 2. Gene expression patterns in the GUDMAP database. (A) Web shot of a WISH basis of a detailed picture of gene expres- screen page in the GUDMAP database. (B) WISH analysis of mammalian transcriptional sion in the genitourinary system. Pages regulators in embryonic day 15.5 UGT. Four images from the database reflect distinct linked to the database will list congenital regional domains in the expression of the indicated regulators (Jing Yu, unpublished data). genitourinary diseases and associated

670 Journal of the American Society of Nephrology J Am Soc Nephrol 19: 667–671, 2008 www.jasn.org SPECIAL ARTICLE genes with expression data from the rius and Jing Yu for the preparation of figures NR, Schaffnit K, Shapovalova NV, Sivisay T, project. Other pages will list details of in this article. Slaughterbeck CR, Smith SC, Smith KA, Smith BI, Sodt AJ, Stewart NN, Stumpf KR, transgenic mice generated by the project Sunkin SM, Sutram M, Tam A, Teemer CD, and their preliminary phenotypic char- Thaller C, Thompson CL, Varnam LR, Visel A, acterization. DISCLOSURES Whitlock RM, Wohnoutka PE, Wolkey CK, In addition to these functions, the da- A.P.M. is a consultant for Merck. Wong VY, Wood M, Yaylaoglu MB, Young tabase will provide the outlet for bioin- RC, Youngstrom BL, Yuan XF, Zhang B, Zwingman TA, Jones AR: Genome-wide at- formatic studies that will maximize the las of gene expression in the adult mouse information and knowledge gained from REFERENCES brain. Nature 445: 168–176, 2007 available GUDMAP data. The key bioin- 6. Challen GA, Martinez G, Davis MJ, Taylor formatic challenge for the consortium 1. Hoy WE, Hughson MD, Bertram JF, Doug- DF, Crowe M, Teasdale RD, Grimmond SM, will be to interrogate, integrate, and mine las-Denton R, Amann K: Nephron number, Little MH: Identifying the molecular pheno- hypertension, renal disease, and renal fail- type of renal progenitor cells. JAmSoc all data sets to identify the best unique ure. J Am Soc Nephrol 16: 2557–2564, Nephrol 15: 2344–2357, 2004 molecular descriptor for every known 2005 7. Martinez G, Georgas K, Challen GA, Rum- anatomic component within the UGT. 2. Little MH, Brennan J, Georgas K, Davies JA, balle B, Davis MJ, Taylor D, Teasdale RD, We also expect to define new molecular Davidson DR, Baldock RA, Beverdam A, Ber- Grimmond SM, Little MH: Definition and subcompartments in the process. This tram JF, Capel B, Chiu HS, Clements D, spatial annotation of the dynamic secretome Cullen-McEwen L, Fleming J, Gilbert T, Her- during early kidney development. Dev Dyn will in turn inform the biology and direct zlinger D, Houghton D, Kaufman MH, Kley- 235: 1709–1719, 2006 decision making with regard to develop- menova E, Koopman PA, Lewis AG, McMa- 8. Schwab K, Patterson LT, Aronow BJ, Luckas ment of further animal tools in this field. hon AP, Mendelsohn CL, Mitchell EK, R, Liang HC, Potter SS: A catalogue of gene We anticipate that the availability of the Rumballe BA, Sweeney DE, Valerius MT, expression in the developing kidney. Kidney data to the community will result in the Yamada G, Yang Y, Yu J: A high-resolution Int 64: 1588–1604, 2003 anatomical ontology of the developing mu- 9. Caruana G, Cullen-McEwen L, Nelson AL, identification of synexpression groups, rine genitourinary tract. Gene Expr Patterns Kostoulias X, Woods K, Gardiner B, Davis whereby coexpression of genes will assist 7: 680–699, 2007 MJ, Taylor DF, Teasdale RD, Grimmond the elucidation of transcriptional regula- 3. Gray PA, Fu H, Luo P, Zhao Q, Yu J, Ferrari SM, Little MH, Bertram JF: Spatial gene tory networks. A, Tenzen T, Yuk DI, Tsung EF, Cai Z, expression in the T-stage mouse metane- The consortium data will also enable Alberta JA, Cheng LP, Liu Y, Stenman JM, phros. Gene Expr Patterns 6: 807–825, Valerius MT, Billings N, Kim HA, Green- 2006 meta-analyses, for example, comparison berg ME, McMahon AP, Rowitch DH, Stiles 10. Mitchell EK, Taylor DF, Woods K, Davis MJ, of UGT development with development CD, Ma Q: Mouse brain organization re- Nelson AL, Teasdale RD, Grimmond SM, Lit- of other organ systems to reveal shared vealed through direct genome-scale TF tle MH, Bertram JF, Caruana G: Differential regulatory principles, with mouse mod- expression analysis. Science 306: 2255– gene expression in the developing mouse els of disease and dysgenesis of the UGT 2257, 2004 ureter. Gene Expr Patterns 6: 519–538, 4. Challen G, Gardiner B, Caruana G, Kostou- 2006 to explore mechanism and pathologies, lias X, Martinez G, Crowe M, Taylor DF, Ber- 11. Potter SS, Hartman HA, Kwan KM, Behringer and eventually intersecting murine and tram J, Little M, Grimmond SM: Temporal RR, Patterson LT: Laser capture-microarray human data to inform better on patient and spatial transcriptional programs in mu- analysis of Lim1 mutant kidney develop- material. Finally, the protocols, tutorials, rine kidney development. Physiol Genomics ment. Genesis 45: 432–439, 2007 links, and resources on the GUDMAP 23: 159–171, 2005 12. Stark K, Vainio S, Vassileva G, McMahon AP: 5. Lein ES, Hawrylycz MJ, Ao N, Ayres M, Bens- Epithelial transformation of metanephric Web site should make this information inger A, Bernard A, Boe AF, Boguski MS, mesenchyme in the developing kidney reg- base an important “one-stop shop” for Brockway KS, Byrnes EJ, Chen L, Chen L, ulated by Wnt-4. Nature 372: 679–683, researchers exploring the UGT. In keep- Chen TM, Chin MC, Chong J, Crook BE, 1994 ing with this idea, efforts are under way Czaplinska A, Dang CN, Datta S, Dee NR, 13. Valerius MT, McMahon AP: Transcriptional to expand the coverage of this site, for Desaki AL, Desta T, Diep E, Dolbeare TA, profiling of Wnt4 mutant mouse kidneys Donelan MJ, Dong HW, Dougherty JG, identifies genes expressed during nephron example, synthesizing information on Duncan BJ, Ebbert AJ, Eichele G, Estin LK, formation. Gene Expr Patterns 2008, in human genetic diseases of the UGT and Faber C, Facer BA, Fields R, Fischer SR, Fliss press highlighting relevant mouse models. TP, Frensley C, Gates SN, Glattfelder KJ, 14. Gong S, Zheng C, Doughty ML, Losos K, Halverson KR, Hart MR, Hohmann JG, How- Didkovsky N, Schambra UB, Nowak NJ, Joy- ell MP, Jeung DP, Johnson RA, Karr PT, ner A, Leblanc G, Hatten ME, Heintz N: A Kawal R, Kidney JM, Knapik RH, Kuan CL, gene expression atlas of the central nervous ACKNOWLEDGMENTS Lake JH, Laramee AR, Larsen KD, Lau C, system based on bacterial artificial chromo- Lemon TA, Liang AJ, Liu Y, Luong LT, somes. Nature 425: 917–925, 2003 We acknowledge the tremendous group ef- Michaels J, Morgan JJ, Morgan RJ, Mortrud 15. Gagneten S, Le Y, Miller J, Sauer B: Brief MT, Mosqueda NF, Ng LL, Ng R, Orta GJ, expression of a GFP cre fusion gene in em- fort of many individual researchers in their Overly CC, Pak TH, Parry SE, Pathak SD, bryonic stem cells allows rapid retrieval of own laboratories toward the GUDMAP goals. Pearson OC, Puchalski RB, Riley ZL, Rockett site-specific genomic deletions. Nucleic Ac- A.P.M. is particularly grateful to Todd Vale- HR, Rowland SA, Royall JJ, Ruiz MJ, Sarno ids Res 25: 3326–3331, 1997

J Am Soc Nephrol 19: 667–671, 2008 Genitourinary Developmental Molecular Anatomy Project 671