Abstract
Haplotype-based methods offer a powerful approach to disease gene mapping, based on the association between causal mutations and the ancestral haplotypes on which they arose. As part of The SNP Consortium Allele Frequency Projects, we characterized haplotype patterns across 51 autosomal regions (spanning 13 megabases of the human genome) in samples from Africa, Europe, and Asia. We show that the human genome can be parsed objectively into haplotype blocks: sizable regions over which there is little evidence for historical recombination and within which only a few common haplotypes are observed. The boundaries of blocks and specific haplotypes they contain are highly correlated across populations. We demonstrate that such haplotype frameworks provide substantial statistical power in association studies of common genetic variation across each region. Our results provide a foundation for the construction of a haplotype map of the human genome, facilitating comprehensive genetic association studies of human disease.
Bibliography
Gabriel, S. B., Schaffner, S. F., Nguyen, H., Moore, J. M., Roy, J., Blumenstiel, B., Higgins, J., DeFelice, M., Lochner, A., Faggart, M., Liu-Cordero, S. N., Rotimi, C., Adeyemo, A., Cooper, R., Ward, R., Lander, E. S., Daly, M. J., & Altshuler, D. (2002). The Structure of Haplotype Blocks in the Human Genome. Science, 296(5576), 2225â2229.
Authors
18
- Stacey B. Gabriel (first)
- Stephen F. Schaffner (additional)
- Huy Nguyen (additional)
- Jamie M. Moore (additional)
- Jessica Roy (additional)
- Brendan Blumenstiel (additional)
- John Higgins (additional)
- Matthew DeFelice (additional)
- Amy Lochner (additional)
- Maura Faggart (additional)
- Shau Neen Liu-Cordero (additional)
- Charles Rotimi (additional)
- Adebowale Adeyemo (additional)
- Richard Cooper (additional)
- Ryk Ward (additional)
- Eric S. Lander (additional)
- Mark J. Daly (additional)
- David Altshuler (additional)
References
43
Referenced
4,633
10.1126/science.274.5287.536
10.1126/science.278.5343.1580
10.1126/science.273.5281.1516
- R. Sachidanandam et al. Nature 409 928 (2001).
-
J. C. Venter et al. Science 291 1304 (2001).
(
10.1556/AVet.49.2001.3.5
) 10.1038/85776
-
E. G. Puffenberger et al. Cell 79 1257 (1994).
(
10.1016/0092-8674(94)90016-7
) -
B. Kerem et al. Science 245 1073 (1989).
(
10.1126/science.2570460
) - J. Hastbacka et al. Nature Genet. 2 204 (1992).
-
J. D. Rioux et al. Nature Genet. 29 223 (2001).
(
10.1016/S0301-0546(01)79061-1
) -
J. P. Hugot et al. Nature 411 599 (2001).
(
10.1038/35079107
) -
Y. Ogura et al. Nature 411 603 (2001).
(
10.1038/35079114
) 10.1086/321275
10.1101/gr.144500
10.1038/76980
10.1038/ng1001-229
10.1038/ng1001-217
- N. Patil et al. Science 294 1719 (2001).
- G. C. Johnson et al. Nature Genet. 29 233 (2001).
- Materials and methods are available as supporting material on Science Online.
-
K. Tang et al. Proc. Natl. Acad. Sci. U.S.A. 96 10016 (1999).
(
10.1073/pnas.96.18.10016
) - Though 82% of assays were successful in at least one population genotyping success rates in each population range from 72 to 79%. The difference between these numbers is due to a low rate of laboratory failure in each attempt.
- L. L. Cavalli-Sforza P. Menozzi A. Piazza The History and Geography of Human Genes (Princeton University Press Princeton NJ 1994).
10.1093/genetics/49.1.49
- An upper confidence bound of 0.98 was used instead of 1.0 because even a single observation of a fourth haplotype makes it mathematically impossible for D′ to be consistent with a value of 1.0 though the confidence interval could be arbitrarily close to 1.0.
-
D. E. Reich et al. Nature 411 199 (2001).
(
10.1038/35075590
) - As a further test of the model we simulated the proportion of pairs at a fixed distance (5 kb) that should show evidence of crossing block boundaries (that is show strong evidence of historical recombination). The model predicts these proportions to be 47% (Yoruban and African-American samples) and 27% (European and Asian samples). In the empirical data we observe 42 and 23% similar to these predictions.
- A low rate of genotyping error is critical to obtaining an accurate measure of haplotype diversity and the proportion in common haplotypes. Even a modest (1 to 2%) genotyping error will create a substantial number of false rare haplotypes; for example with a 10-marker haplotype and a 2% error rate 18% of chromosomes will contain at least one error and thus not match the few common haplotypes.
- Within blocks the common haplotypes showed little evidence for historical recombination. For example we performed the four gamete tests using SNPs drawn only from haplotypes with frequency 5% or higher in each block. One or more violations of the four gamete tests were observed in only 5% of the blocks.
- To maximize power these comparisons were made only for SNP pairs spaced 5 to 10 kb apart. At shorter distances nearly all SNP pairs are in a single block; at greater distances most SNP pairs are in different blocks.
- Blocks and haplotypes were identified separately in each population sample and the results were compared for those blocks that were physically overlapping in all three samples.
10.1093/genetics/106.3.479
10.1126/science.3125610
10.1073/pnas.95.14.8119
10.1038/35047064
-
S. A. Tishkoff et al. Science 271 1380 (1996).
(
10.1126/science.271.5254.1380
) - S. A. Tishkoff et al. Am. J. Hum. Genet. 67 901 (2000).
- We examined SNP pairs that were in different blocks in the Yoruban samples but in a single block in the European sample. Such pairs had higher D′ values in the Yoruban sample (D′ = 0.46) than pairs found in different blocks in both population samples (D′ = 0.28). The average frequency of all haplotypes in the Yoruban population was 0.21 whereas those that were found only in the Yoruban sample (but not in the European and Asian samples) had a mean frequency of 0.16.
- A. G. Clark et al. Am. J. Hum. Genet. 63 595 (1998).
- A. R. Templeton et al. Am. J. Hum. Genet. 66 69 (2000).
- S. M. Fullerton et al. Am. J. Hum. Genet. 67 881 (2000).
- The small fraction of SNPs that show r 2 values <0.5 could be attributable to a range of causes: branches of the gene tree not defined with the number of markers used gene conversion events or recurrent mutations. We note that errors in genotyping or map position decrease (but cannot increase) the value of r 2 .
- This work was supported by a grant to D.A. from The SNP Consortium. We thank members of the Program in Medical and Population Genetics at the Whitehead/MIT Center for Genome Research for helpful discussion particularly J. Hirschhorn D. Reich and N. Patterson. The authors also thank the following colleagues for sharing data and analyses before publication: D. Bentley (Sanger Institute); L. Cardon (Oxford); and D. Cutler M. Zwick and A. Chakravarti (Johns Hopkins). D.A. is a Charles E. Culpeper Scholar of the Rockefeller Brothers Fund and a Burroughs Welcome Fund Clinical Scholar in Translational Research. Supporting Online Material www.sciencemag.org/cgi/content/full/1069424/DC1 Materials and Methods Figs. S1 to S3 Table S1
Dates
Type | When |
---|---|
Created | 23 years ago (July 28, 2002, 3:06 p.m.) |
Deposited | 1 year, 7 months ago (Jan. 9, 2024, 5:40 p.m.) |
Indexed | 7 hours, 44 minutes ago (Aug. 23, 2025, 9:45 p.m.) |
Issued | 23 years, 2 months ago (June 21, 2002) |
Published | 23 years, 2 months ago (June 21, 2002) |
Published Print | 23 years, 2 months ago (June 21, 2002) |
@article{Gabriel_2002, title={The Structure of Haplotype Blocks in the Human Genome}, volume={296}, ISSN={1095-9203}, url={http://dx.doi.org/10.1126/science.1069424}, DOI={10.1126/science.1069424}, number={5576}, journal={Science}, publisher={American Association for the Advancement of Science (AAAS)}, author={Gabriel, Stacey B. and Schaffner, Stephen F. and Nguyen, Huy and Moore, Jamie M. and Roy, Jessica and Blumenstiel, Brendan and Higgins, John and DeFelice, Matthew and Lochner, Amy and Faggart, Maura and Liu-Cordero, Shau Neen and Rotimi, Charles and Adeyemo, Adebowale and Cooper, Richard and Ward, Ryk and Lander, Eric S. and Daly, Mark J. and Altshuler, David}, year={2002}, month=jun, pages={2225–2229} }