Disease-associated alleles in genome-wide association studies are enriched for derived low frequency alleles relative to HapMap and neutral expectations

Table 1 Disease-associated alleles vs. null expectations

	Mean frequency of a randomly chosen allele	Proportion ancestral
Null expectations
HapMap data (n = 1000, unweighted)	0.721	0.623
HapMap data (n = 1000, weighted)	0.610	0.568
Theoretical (neutral, unweighted)	0.741	0.741
Theoretical (neutral, weighted)	0.650	0.650
Disease-associated alleles
Cancer (n = 112)	0.362**	0.446*
Cardiovascular (n = 145)	0.364**	0.379**
Metabolism (n = 160)	0.365**	0.456*
Miscellaneous disease (n = 290)	0.413**	0.434**
Morphological (n = 276)	0.409**	0.467*
Neurological (n = 135)	0.429**	0.430*
Multiple phenotypic classes (n = 25)	0.312*	0.320*
All GWAS alleles (n = 1143)	0.394**	0.437**
All replicated GWAS alleles (n = 142)	0.396**	0.437**

Unweighted values do not incorporate the probability of detection, and weighted values incoporate the probability of detection in a GWAS. Neutral expectations use a polymorphism threshold (d) of 0.025 and Equations 5 and 10. Relative magnitudes of disease-associated and control allele frequencies were compared via Mann-Whitney U tests, and proportions of ancestral alleles were compared via binomial tests. * indicates significant differences from HapMap and neutral scenarios (p-value < 0.05), and ** indicates highly significant differences from HapMap and neutral scenarios (p-value < 0.0001). Totals for GWAS alleles are labeled in boldface type. Replicated alleles are those alleles that have been implicated in multiple studies.

ISSN: 1755-8794