A high-density admixture map for disease gene discovery in african americans - PubMed (original) (raw)

. 2004 May;74(5):1001-13.

doi: 10.1086/420856. Epub 2004 Apr 14.

Nick Patterson, James A Lautenberger, Ann L Truelove, Gavin J McDonald, Alicja Waliszewska, Bailey D Kessing, Michael J Malasky, Charles Scafe, Ernest Le, Philip L De Jager, Andre A Mignault, Zeng Yi, Guy De The, Myron Essex, Jean-Louis Sankale, Jason H Moore, Kwabena Poku, John P Phair, James J Goedert, David Vlahov, Scott M Williams, Sarah A Tishkoff, Cheryl A Winkler, Francisco M De La Vega, Trevor Woodage, John J Sninsky, David A Hafler, David Altshuler, Dennis A Gilbert, Stephen J O'Brien, David Reich

Affiliations

A high-density admixture map for disease gene discovery in african americans

Michael W Smith et al. Am J Hum Genet. 2004 May.

Abstract

Admixture mapping (also known as "mapping by admixture linkage disequilibrium," or MALD) provides a way of localizing genes that cause disease, in admixed ethnic groups such as African Americans, with approximately 100 times fewer markers than are required for whole-genome haplotype scans. However, it has not been possible to perform powerful scans with admixture mapping because the method requires a dense map of validated markers known to have large frequency differences between Europeans and Africans. To create such a map, we screened through databases containing approximately 450000 single-nucleotide polymorphisms (SNPs) for which frequencies had been estimated in African and European population samples. We experimentally confirmed the frequencies of the most promising SNPs in a multiethnic panel of unrelated samples and identified 3011 as a MALD map (1.2 cM average spacing). We estimate that this map is approximately 70% informative in differentiating African versus European origins of chromosomal segments. This map provides a practical and powerful tool, which is freely available without restriction, for screening for disease genes in African American patient cohorts. The map is especially appropriate for those diseases that differ in incidence between the parental African and European populations.

PubMed Disclaimer

Figures

Figure 1

Figure 1

Power of the individual 2,154 MALD markers (white bars), as a fraction of the maximum (power is calculated by eq. [1]). Power increases strikingly when the methods described by Patterson et al. ( [in this issue]) are used to combine data from multiple, closely linked markers (black bars).

Figure 2

Figure 2

Power of the 2,154-marker map as a fraction of what would be expected if there was full information about ancestry available at every point (e.g., nearly complete information [a value of 1 on the _Y_-axis] is available at Duffy on chromosome 1, indicated by a star). The black line is the power of the African American map under the assumption of an average of six generations since admixture (the power would be lower for individuals with more generations since admixture). We were also able to roughly estimate the quality of the map for studies involving Amerindian and East Asian mixtures under the assumption of six generations since admixture, although these estimates are less reliable because the sample sizes were smaller.

Figure 3

Figure 3

A, Distribution of percentage of European ancestry (M i) in 109 African American samples genotyped at 2,154 MALD markers (mean ± SD = 20% ± 8%). B, Distribution of estimated number of generations since admixture (λ_i_) for 109 African Americans (mean ± SD = 6.3 ± 1.1). All estimates are generated by the ANCESTRYMAP software from the accompanying article by Patterson et al. ( [in this issue]). We emphasize that the number of “generations since admixture” is an average across a person’s lineages, since mixing between Africans and Europeans has occurred over many generations.

Figure 4

Figure 4

Predicted probability of no recombination having occurred since admixture between a disease locus and a nearby marker locus, predicting the shape of a peak of association in an African American admixture study. The analysis was based on the distribution of the estimated number of generations since admixture (λ_i_) values estimated from 109 African Americans (fig. 3_B_). We present the probability of a crossover between ancestry segments, as a function of the distance between a disease locus and mapping marker. At 4.5 cM, we expect the strength of association to drop to 75% of its maximum (measured as LOD or log of the P value). At 11 cM and 23 cM, the power drops to 50% and 25% of its maximum, respectively.

Figure 5

Figure 5

Estimates of ancestry along chromosome 1 for five African American samples, estimated by use of the Patterson et al. ( [in this issue]) ANCESTRYMAP software. The positions of the 169 markers in the map on chromosome 1 that were used for this inference are indicated by hash marks (inferred European chromosome segments are indicated by black bars). Sharp transitions between segments of 0, 1, or 2 European alleles are clear from the analysis, indicating the high resolution of the map.

Similar articles

Cited by

References

Electronic-Database Information

    1. Applied Biosystems myScience Web site, http://myscience.appliedbiosystems.com
    1. dbSNP, http://www.ncbi.nlm.nih.gov/SNP/
    1. Laboratory of Genomic Diversity, http://home.ncifcrf.gov/ccr/lgd/human_genome/mald/index_n.asp (for figures showing positions of map SNPs in the genome)
    1. SNP Consortium, http://snp.cshl.org/allele_frequency_project/
    1. University of Washington Fred Hutchinson Cancer Research Center (UW-FHCRC) Variation Discovery Resource, http://pga.gs.washington.edu/

References

    1. Adams MD, Cargill MA, Spier EG, De La Vega FM, Olson SJ, White TJ, Sninsky JJ, Gilbert DA, Hunkapiller MW (2002) Applied genomics: exploring functional variation and gene expression. Am J Hum Genet Suppl 71:203
    1. Briscoe D, Stephens JC, O’Brien SJ (1994) Linkage disequilibrium in admixed populations: applications in gene mapping. J Hered 85:59–63 - PubMed
    1. Carlson CS, Eberle MA, Rieder MJ, Smith JD, Kruglyak L, Nickerson DA (2003) Additional SNPs and linkage-disequilibrium analyses are necessary for whole-genome association studies in humans. Nat Genet 33:518–52110.1038/ng1128 - DOI - PubMed
    1. Chakraborty R, Weiss KM (1988) Admixture as a tool for finding linked genes and detecting that difference from allelic association between loci. Proc Natl Acad Sci USA 85:9119–9123 - PMC - PubMed
    1. Chen M-H, Shao Q-M, Ibrahim JG (2000) Monte Carlo methods in Bayesian computation. Springer, New York

Publication types

MeSH terms

Substances

Grants and funding

LinkOut - more resources