Exploring the metabolic and genetic control of gene expression on a genomic scale

Science

Volumn 278, Issue 5338, 1997, Pages 680-686

  DeRisi, Joseph L ^a   Iyer, Vishwanath R ^a   Brown, Patrick O ^a  

^a STANFORD UNIVERSITY SCHOOL OF MEDICINE   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

COMPLEMENTARY DNA; DNA BINDING PROTEIN; MESSENGER RNA; RIBOSOME PROTEIN;

ARTICLE; ENERGY METABOLISM; GENE EXPRESSION REGULATION; GENETIC TRANSCRIPTION; GENOME; METABOLIC REGULATION; NONHUMAN; OPEN READING FRAME; POLYMERASE CHAIN REACTION; PRIORITY JOURNAL; PROTEIN SYNTHESIS; SACCHAROMYCES CEREVISIAE;

CITRIC ACID CYCLE; CULTURE MEDIA; DNA-BINDING PROTEINS; FERMENTATION; FUNGAL PROTEINS; GENE EXPRESSION REGULATION, FUNGAL; GENES, FUNGAL; GENES, REGULATOR; GENOME, FUNGAL; GLUCONEOGENESIS; GLUCOSE; GLYOXYLATES; NUCLEAR PROTEINS; OPEN READING FRAMES; OXYGEN CONSUMPTION; REPRESSOR PROTEINS; RNA, FUNGAL; RNA, MESSENGER; SACCHAROMYCES CEREVISIAE; SACCHAROMYCES CEREVISIAE PROTEINS; TRANSCRIPTION FACTORS;

SACCHAROMYCES CEREVISIAE;

EID: 0030669030     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.278.5338.680     Document Type: Article

References (59)

1. M. Schena, D. Shalon, R. W. Davis, P. O. Brown, Science 270, 467 (1995).
2. D. Shalon, S. J. Smith, P. O. Brown, Genome Res. 6, 639 (1996).
3. D. Lashkari, Proc. Natl. Acad. Sci. U.S.A., in press.
4. J. DeRisi et al., Nature Genet. 14, 457 (1996).
5. D. J. Lockhart et al., Nature Biotechnol. 14, 1675 (1996).
6. M. Chee et al., Science 274, 610 (1996).
7. M. Johnston and M. Carlson, in The Molecular Biology of the Yeast Saccharomyces: Gene Expression, E. W. Jones, J. R. Pringle, J. R. Broach, Eds. (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1992), p. 193.
8. Primers for each known or predicted protein coding sequence were supplied by Research Genetics. PCR was performed with the protocol supplied by Research Genetics, using genomic DNA from yeast strain S288C as a template. Each PCR product was verified by agarose gel electrophoresis and was deemed correct if the lane contained a single band of appropriate mobility. Failures were marked as such in the database. The overall success rate for a single-pass amplification of 6116 ORFs was ∼94.5%.
9. Glass slides (Gold Seal) were cleaned for 2 hours in a solution of 2 N NaOH and 70% ethanol. After rinsing in distilled water, the slides were then treated with a 1:5 dilution of poly-L-lysine adhesive solution (Sigma) for 1 hour, and then dried for 5 min at 40°C in a vacuum oven. DNA samples from 100-μl PCR reactions were purified by ethanol purification in 96-well microtiter plates. The resulting precipitates were resuspended in 3x standard saline citrate (SSC) and transferred to new plates for arraying. A custom-built arraying robot was used to print on a batch of 110 slides. Details of the design of the microarrayer are available at cmgm.stanford.edu/pbrown. After printing, the microarrays were rehydrated for 30 s in a humid chamber and then snap-dried for 2 s on a hot plate (100°C). The DNA was then ultraviolet (UV)-crosslinked to the surface by subjecting the slides to 60 mJ of energy (Stratagene Stratalinker). The rest of the poly-L-lysine surface was blocked by a 15-min incubation in a solution of 70 mM succinic anhydride dissolved in a solution consisting of 315 ml of 1-methyl-2-pyrrolidinone (Aldrich) and 35 ml of 1 M boric acid (pH 8.0). Directly after the blocking reaction, the bound DNA was denatured by a 2-min incubation in distilled water at ∼95°C. The slides were then transferred into a bath of 100% ethanol at room temperature, rinsed, and then spun dry in a clinical centrifuge. Slides were stored in a closed box at room temperature until used.
10. YPD medium (8 liters), in a 10-liter fermentation vessel, was inoculated with 2 ml of a fresh overnight culture of yeast strain DBY7286 (MATa, ura3, GAL2). The fermentor was maintained at 30°C with constant agitation and aeration. The glucose content of the media was measured with a UV test kit (Boehringer Mannheim, catalog number 716251) Cell density was measured by OD at 600-nm wavelength. Aliquots of culture were rapidly withdrawn from the fermentation vessel by peristaltic pump, spun down at room temperature, and then flash frozen with liquid nitrogen. Frozen cells were stored at -80°C.
11. +] RNA, primed by a dT(16) oligomer. This mixture was heated to 70°C for 10 min, and then transferred to ice. A premixed solution, consisting of 200 U Superscript II (Gibco), buffer, deoxyribonucleoside triphosphates, and fluorescent nucleotides, was added to the RNA. Nucleotides were used at these final concentrations: 500 μM for dATP, dCTP, and dGTP and 200 μM for dTTP. Cy3-dUTP and Cy5-dUTP were used at a final concentration of 100 μM. The reaction was then incubated at 42°C for 2 hours. Unincorporated fluorescent nucleotides were removed by first diluting the reaction mixture with of 470 μl of 10 mM tris-HCl (pH 8.0)/1 mM EDTA and then subsequently concentrating the mix to ∼5 μl, using Centricon-30 microconcentrators (Amicon).
12. Purified, labeled cDNA was resuspended in 11 μl of 3.5 x SSC containing 10 μg poly(dA) and 0.3 μl of 10% SDS. Before hybridization, the solution was boiled for 2 min and then allowed to cool to room temperature. The solution was applied to the microarray under a cover slip, and the slide was placed in a custom hybridization chamber which was subsequently incubated for ∼8 to 12 hours in a water bath at 62°C. Before scanning, slides were washed in 2x SSC, 0.2% SDS for 5 min, and then 0.05x SSC for 1 min. Slides were dried before scanning by centrifugation at 500 rpm in a Beckman CS-6R centrifuge.
13. The complete data set is available on the Internet at cmgm.stanford.edu/pbrown/explore/index.html
14. For 95% of all the genes analyzed, the mRNA levels measured in cells harvested at the first and second interval after inoculation differed by a factor of less than 1.5. The correlation coefficient for the comparison between mRNA levels measured for each gene in these two different mRNA samples was 0.98. When duplicate mRNA preparations from the same cell sample were compared in the same way, the correlation coefficient between the expression levels measured for the two samples by comparative hybridization was 0.99.
15. The numbers and identities of known and putative genes, and their homologies to other genes, were gathered from the following public databases: Saccharomyces Genome Database (genome-www. stanford.edu), Yeast Protein Database (quest7. proteome.com), and Munich Information Centre for Protein Sequences (speedy.mips.biochem.mpg.de/ mips/yeast/index.htmlx).
16. A. Scholer and H. J. Schuller, Mol. Cell. Biol. 14, 3613 (1994).
17. S. Kratzer and H. J. Schuller, Gene 161, 75 (1995).
18. R. J. Haselbeck and H. L. McAlister, J. Biol. Chem. 268, 12116 (1993).
19. M. Fernandez, E. Fernandez, R. Rodicio, Mol. Gen. Genet. 242, 727 (1994).
20. A. Hartig et al., Nucleic Acids Res. 20, 5677 (1992).
21. P. M. Martinez et al., EMBO J. 15, 2227 (1996).
22. J. C. Varela, U. M. Praekelt, P. A. Meacock, R. J. Planta, W. H. Mager, Mol. Cell. Biol. 15, 6232 (1995).
23. H. Ruis and C. Schuller, Bioessays 17, 959 (1995).
24. J. L. Parrou, M. A. Teste, J. Francois, Microbiology 143, 1891 (1997).
25. This expression profile was defined as having an induction of greater than 10-fold at 18.5 hours and less than 11-fold at 20.5 hours.
26. S. L. Forsburg and L. Guarente, Genes Dev. 3, 1166 (1989).
27. J. T. Olesen and L. Guarente, ibid. 4, 1714 (1990).
28. M. Rosenkrantz, C. S. Kell, E. A. Pennell, L. J. Devenish, Mol. Microbiol. 13, 119 (1994).
29. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr. The nucleotide codes are as follows: B-C, G, or T; N-G, A, T, or C; R-A or G; and Y-C or T.
30. C. Fondrat and A. Kalogeropoulos, Comput. Appl. Biosci. 12, 363 (1996).
31. D. Shore, Trends Genet. 10, 408 (1994).
32. R. J. Planta and H. A. Raue, ibid. 4, 64 (1988).
33. The degenerate consensus sequence VYCYRNNC-MNH was used to search for potential RAP1-binding sites. The exact consensus, as defined by (30), is WACAYCCRTACATYW, with up to three differences allowed.
34. S. F. Neuman, S. Bhattacharya, J. R. Broach, Mol. Cell. Biol. 15, 3187 (1995).
35. P. Lesage, X. Yang, M. Carlson, ibid. 16, 1921 (1996).
36. For example, we observed large inductions of the genes coding for PCK1, FBP1 [Z. Yin et al., Mol. Microbiol. 20, 751 (1996)], the central glyoxylate cycle gene ICL1 [A. Scholer and H. J. Schuller, Curr. Genet. 23, 375 (1993)], and the "aerobic" isoform of acetyl-CoA synthase, ACS1 [M. A. van den Berg et al., J. Biol. Chem. 271, 28953 (1996)], with concomitant down-regulation of the glycolytic-specific genes PYK1 and PFK2 [P. A. Moore et al., Mol. Cell. Biol. 11, 5330 (1991)]. Other genes not directly involved in carbon metabolism but known to be induced upon nutrient limitation include genes encoding cytosolic catalase T CTT1 [P. H. Bissinger et al., ibid. 9, 1309 (1989)] and several genes encoding small heat-shock proteins, such as HSP12, HSP26, and HSP42 [I. Farkas et al., J. Biol. Chem. 266, 15602 (1991); U. M. Praekelt and P. A. Meacock, Mol. Gen. Genet. 223, 97 (1990); D. Wotton et al., J. Biol. Chem. 271, 2717 (1996)].
37. For example, we observed large inductions of the genes coding for PCK1, FBP1 [Z. Yin et al., Mol. Microbiol. 20, 751 (1996)], the central glyoxylate cycle gene ICL1 [A. Scholer and H. J. Schuller, Curr. Genet. 23, 375 (1993)], and the "aerobic" isoform of acetyl-CoA synthase, ACS1 [M. A. van den Berg et al., J. Biol. Chem. 271, 28953 (1996)], with concomitant down-regulation of the glycolytic-specific genes PYK1 and PFK2 [P. A. Moore et al., Mol. Cell. Biol. 11, 5330 (1991)]. Other genes not directly involved in carbon metabolism but known to be induced upon nutrient limitation include genes encoding cytosolic catalase T CTT1 [P. H. Bissinger et al., ibid. 9, 1309 (1989)] and several genes encoding small heat-shock proteins, such as HSP12, HSP26, and HSP42 [I. Farkas et al., J. Biol. Chem. 266, 15602 (1991); U. M. Praekelt and P. A. Meacock, Mol. Gen. Genet. 223, 97 (1990); D. Wotton et al., J. Biol. Chem. 271, 2717 (1996)].
38. For example, we observed large inductions of the genes coding for PCK1, FBP1 [Z. Yin et al., Mol. Microbiol. 20, 751 (1996)], the central glyoxylate cycle gene ICL1 [A. Scholer and H. J. Schuller, Curr. Genet. 23, 375 (1993)], and the "aerobic" isoform of acetyl-CoA synthase, ACS1 [M. A. van den Berg et al., J. Biol. Chem. 271, 28953 (1996)], with concomitant down-regulation of the glycolytic-specific genes PYK1 and PFK2 [P. A. Moore et al., Mol. Cell. Biol. 11, 5330 (1991)]. Other genes not directly involved in carbon metabolism but known to be induced upon nutrient limitation include genes encoding cytosolic catalase T CTT1 [P. H. Bissinger et al., ibid. 9, 1309 (1989)] and several genes encoding small heat-shock proteins, such as HSP12, HSP26, and HSP42 [I. Farkas et al., J. Biol. Chem. 266, 15602 (1991); U. M. Praekelt and P. A. Meacock, Mol. Gen. Genet. 223, 97 (1990); D. Wotton et al., J. Biol. Chem. 271, 2717 (1996)].
39. For example, we observed large inductions of the genes coding for PCK1, FBP1 [Z. Yin et al., Mol. Microbiol. 20, 751 (1996)], the central glyoxylate cycle gene ICL1 [A. Scholer and H. J. Schuller, Curr. Genet. 23, 375 (1993)], and the "aerobic" isoform of acetyl-CoA synthase, ACS1 [M. A. van den Berg et al., J. Biol. Chem. 271, 28953 (1996)], with concomitant down-regulation of the glycolytic-specific genes PYK1 and PFK2 [P. A. Moore et al., Mol. Cell. Biol. 11, 5330 (1991)]. Other genes not directly involved in carbon metabolism but known to be induced upon nutrient limitation include genes encoding cytosolic catalase T CTT1 [P. H. Bissinger et al., ibid. 9, 1309 (1989)] and several genes encoding small heat-shock proteins, such as HSP12, HSP26, and HSP42 [I. Farkas et al., J. Biol. Chem. 266, 15602 (1991); U. M. Praekelt and P. A. Meacock, Mol. Gen. Genet. 223, 97 (1990); D. Wotton et al., J. Biol. Chem. 271, 2717 (1996)].
40. For example, we observed large inductions of the genes coding for PCK1, FBP1 [Z. Yin et al., Mol. Microbiol. 20, 751 (1996)], the central glyoxylate cycle gene ICL1 [A. Scholer and H. J. Schuller, Curr. Genet. 23, 375 (1993)], and the "aerobic" isoform of acetyl-CoA synthase, ACS1 [M. A. van den Berg et al., J. Biol. Chem. 271, 28953 (1996)], with concomitant down-regulation of the glycolytic-specific genes PYK1 and PFK2 [P. A. Moore et al., Mol. Cell. Biol. 11, 5330 (1991)]. Other genes not directly involved in carbon metabolism but known to be induced upon nutrient limitation include genes encoding cytosolic catalase T CTT1 [P. H. Bissinger et al., ibid. 9, 1309 (1989)] and several genes encoding small heat-shock proteins, such as HSP12, HSP26, and HSP42 [I. Farkas et al., J. Biol. Chem. 266, 15602 (1991); U. M. Praekelt and P. A. Meacock, Mol. Gen. Genet. 223, 97 (1990); D. Wotton et al., J. Biol. Chem. 271, 2717 (1996)].
41. For example, we observed large inductions of the genes coding for PCK1, FBP1 [Z. Yin et al., Mol. Microbiol. 20, 751 (1996)], the central glyoxylate cycle gene ICL1 [A. Scholer and H. J. Schuller, Curr. Genet. 23, 375 (1993)], and the "aerobic" isoform of acetyl-CoA synthase, ACS1 [M. A. van den Berg et al., J. Biol. Chem. 271, 28953 (1996)], with concomitant down-regulation of the glycolytic-specific genes PYK1 and PFK2 [P. A. Moore et al., Mol. Cell. Biol. 11, 5330 (1991)]. Other genes not directly involved in carbon metabolism but known to be induced upon nutrient limitation include genes encoding cytosolic catalase T CTT1 [P. H. Bissinger et al., ibid. 9, 1309 (1989)] and several genes encoding small heat-shock proteins, such as HSP12, HSP26, and HSP42 [I. Farkas et al., J. Biol. Chem. 266, 15602 (1991); U. M. Praekelt and P. A. Meacock, Mol. Gen. Genet. 223, 97 (1990); D. Wotton et al., J. Biol. Chem. 271, 2717 (1996)].
42. For example, we observed large inductions of the genes coding for PCK1, FBP1 [Z. Yin et al., Mol. Microbiol. 20, 751 (1996)], the central glyoxylate cycle gene ICL1 [A. Scholer and H. J. Schuller, Curr. Genet. 23, 375 (1993)], and the "aerobic" isoform of acetyl-CoA synthase, ACS1 [M. A. van den Berg et al., J. Biol. Chem. 271, 28953 (1996)], with concomitant down-regulation of the glycolytic-specific genes PYK1 and PFK2 [P. A. Moore et al., Mol. Cell. Biol. 11, 5330 (1991)]. Other genes not directly involved in carbon metabolism but known to be induced upon nutrient limitation include genes encoding cytosolic catalase T CTT1 [P. H. Bissinger et al., ibid. 9, 1309 (1989)] and several genes encoding small heat-shock proteins, such as HSP12, HSP26, and HSP42 [I. Farkas et al., J. Biol. Chem. 266, 15602 (1991); U. M. Praekelt and P. A. Meacock, Mol. Gen. Genet. 223, 97 (1990); D. Wotton et al., J. Biol. Chem. 271, 2717 (1996)].
43. For example, we observed large inductions of the genes coding for PCK1, FBP1 [Z. Yin et al., Mol. Microbiol. 20, 751 (1996)], the central glyoxylate cycle gene ICL1 [A. Scholer and H. J. Schuller, Curr. Genet. 23, 375 (1993)], and the "aerobic" isoform of acetyl-CoA synthase, ACS1 [M. A. van den Berg et al., J. Biol. Chem. 271, 28953 (1996)], with concomitant down-regulation of the glycolytic-specific genes PYK1 and PFK2 [P. A. Moore et al., Mol. Cell. Biol. 11, 5330 (1991)]. Other genes not directly involved in carbon metabolism but known to be induced upon nutrient limitation include genes encoding cytosolic catalase T CTT1 [P. H. Bissinger et al., ibid. 9, 1309 (1989)] and several genes encoding small heat-shock proteins, such as HSP12, HSP26, and HSP42 [I. Farkas et al., J. Biol. Chem. 266, 15602 (1991); U. M. Praekelt and P. A. Meacock, Mol. Gen. Genet. 223, 97 (1990); D. Wotton et al., J. Biol. Chem. 271, 2717 (1996)].
44. The levels of induction we measured for genes that were expressed at very low levels in the uninduced state (notably, FBP1 and PCK1) were generally lower than those previously reported. This discrepancy was likely due to the conservative background subtraction method we used, which generally resulted in overestimation of very low expression levels (46).
45. Cross-hybridization of highly related sequences can also occasionally obscure changes in gene expression, an important concern where members of gene families are functionally specialized and differentially regulated. The major alcohol dehydrogenase genes, ADH1 and ADH2, share 88% nucleotide identity. Reciprocal regulation of these genes is an important feature of the diauxic shift, but was not observed in this experiment, presumably because of cross-hybridization of the fluorescent cDNAs representing these two genes. Nevertheless, we were able to detect differential expression of closely related isoforms of other enzymes, such as HXK1/HXK2 (77% identical) [P. Herrero et al., Yeast 11, 137 (1995)], MLS1/ DAL7 (73% identical) (20), and PGM1/PGM2 (72% identical) [D. Oh, J. E. Hopper, Mol. Cell. Biol. 10, 1415 (1990)], in accord with previous studies. Use in the microarray of deliberately selected DNA sequences corresponding to the most divergent segments of homologous genes, in lieu of the complete gene sequences, should relieve this problem in many cases.
46. Cross-hybridization of highly related sequences can also occasionally obscure changes in gene expression, an important concern where members of gene families are functionally specialized and differentially regulated. The major alcohol dehydrogenase genes, ADH1 and ADH2, share 88% nucleotide identity. Reciprocal regulation of these genes is an important feature of the diauxic shift, but was not observed in this experiment, presumably because of cross-hybridization of the fluorescent cDNAs representing these two genes. Nevertheless, we were able to detect differential expression of closely related isoforms of other enzymes, such as HXK1/HXK2 (77% identical) [P. Herrero et al., Yeast 11, 137 (1995)], MLS1/ DAL7 (73% identical) (20), and PGM1/PGM2 (72% identical) [D. Oh, J. E. Hopper, Mol. Cell. Biol. 10, 1415 (1990)], in accord with previous studies. Use in the microarray of deliberately selected DNA sequences corresponding to the most divergent segments of homologous genes, in lieu of the complete gene sequences, should relieve this problem in many cases.
47. F. E. Williams, U. Varanasi, R. J. Trumbly, Mol. Cell. Biol. 11, 3307 (1991).
48. D. Tzamarias and K. Struhl, Nature 369, 758 (1994).
49. Differences in mRNA levels between the tup1Δ and wild-type strain were measured in two independent experiments. The correlation coefficient between the complete sets of expression ratios measured in these duplicate experiments was 0.83. The concordance between the sets of genes that appeared to be induced was very high between the two experiments. When only the 355 genes that showed at least a twofold increase in mRNA in the tup1Δ strain in either of the duplicate experiments were compared, the correlation coefficient was 0.82.
50. The tup1Δ mutation consists of an insertion of the LEU2 coding sequence, including a stop codon, between the ATG of TUP1 and an Eco R I site 124 base pairs before the stop codon of the TUP1 gene.
51. L. R. Kowalski, K. Kondo, M. Inouye, Mol. Microbiol. 15, 341 (1995).
52. M. Viswanathan, G. Muthukumar, Y. S. Cong, J. Lenard, Gene 148, 149 (1994).
53. D. Hirata, K. Yano, T. Miyakawa, Mol. Gen. Genet. 242, 250 (1994).
54. A. Gutierrez, L. Caramelo, A. Prieto, M. J. Martinez, A. T. Martinez, Appl. Environ. Microbiol. 60, 1783 (1994).
55. A. Muheim et al., Eur. J. Biochem. 195, 369 (1991).
56. J. A. Wemmie, M. S. Szczypka, D. J. Thiele, W. S. Moye-Rowley, J. Biol. Chem. 269, 32592 (1994).
57. Microarrays were scanned using a custom-built scanning laser microscope built by S. Smith with software written by N. Ziv. Details concerning scanner design and construction are available at cmgm. stanford.edu/pbrown. Images were scanned at a resolution of 20 μm per pixel. A separate scan, using the appropriate excitation line, was done for each of the two fluorophores used. During the scanning process, the ratio between the signals in the two channels was calculated for several array elements containing total genomic DNA. To normalize the two channels with respect to overall intensity, we then adjusted photomultiplier and laser power settings such that the signal ratio at these elements was as close to 1.0 as possible. The combined images were analyzed with custom-written software. A bounding box, fitted to the size of the DNA spots in each quadrant, was placed over each array element. The average fluorescent intensity was calculated by summing the intensities of each pixel present in a bounding box, and then dividing by the total number of pixels. Local area background was calculated for each array element by determining the average fluorescent intensity for the lower 20% of pixel intensities. Although this method tends to underestimate the background, causing an underestimation of extreme ratios, it produces a very consistent and noise-tolerant approximation. Although the analog-to-digital board used for data collection possesses a wide dynamic range (12 bits), several signals were saturated (greater than the maximum signal intensity allowed) at the chosen settings. Therefore, exteme ratios at bright elements are generally underestimated. A signal was deemed significant if the average intensity after background subtraction was at least 2.5-fold higher than the standard deviation in the background measurements for all elements on the array.
58. In addition to the 17 genes shown in Table 1, three additional genes were induced by an average of more than threefold in the duplicate experiments, but in one of the two experiments, the induction was less than twofold (range 1.6-to 1.9-fold)
59. We thank H. Bennett, P. Spellman, J. Ravetto, M. Eisen, R. Pillai, B. Dunn, T. Ferea, and other members of the Brown lab for their assistance and helpful advice. We also thank S. Friend, D. Botstein, S. Smith, J. Hudson, and D. Dolginow for advice, support, and encouragement; K. Struhl and S. Chatterjee for the Tup1 deletion strain; L. Fernandes for helpful advice on Yap1; and S. Klapholz and the reviewers for many helpful comments on the manuscript. Supported by a grant from the National Human Genome Research Institute (NHGRI) (HG00450), and by the Howard Hughes Medical Institute (HHMI). J.D.R. was supported by the HHMI and the NHGRI. V.R. was supported in part by an Institutional Training Grant in Genome Science (T32 HG00044) from the NHGRI. P.O.B. is an associate investigator of the HHMI.