Table 1.

Admixture association scores on chromosome 1 in different sample sets.

Study	Cases	Controls	Highest LOD score on chr. 1 case-only	Highest LOD score on chr. 1 cases + controls	Highest case–control Z-score on chr. 1	Empirical p-value for admixture association	95% CI for risk per European chromosome
2005 study	605	1043	5.2	5.2	3.3	N/A	1.27–1.70
2007 follow-up	1044	1161	9.2	9.3	4.2	< 0.001	1.32–1.62
Full new cohort	1305	1155	3.9	2.7	3.4	0.114	1.16–1.43
2007 subset of new cohort	899	1155	9.8	9.3	4.5	< 0.001	1.37–1.76
Samples added after 2007	406	1155	− 3.6	− 3.0	0.9	1	0.59–0.91

The highest LOD score results are computed by ANCESTRYMAP based on a prior on relative risk per European ancestry allele of 1.5. The 95% confidence intervals for risk per European allele are obtained by running on a uniformly spaced grid of models from 0.5 to 2.0-fold per European allele, assuming an equal prior probability of risk for each, and then taking the LOD score to the power of 10 and normalizing to obtain a posterior. The LOD scores for the grid of models for the 2005 study are from the original publication¹¹; all LOD scores are given in Supplementary Table S3. Empirical p-values were found through permutation analysis; see “Materials and methods” section. p-values were listed as < 0.001 when 0 of the 1000 permutations had a score at least as large as that of the LOD score of that data subset.