Figure 1.
Relationship of Roma with other worldwide populations.
We applied PCA and ADMIXTURE to study the relationship of Roma with the HapMap and South Asian populations. In PCA, each point represents an individual, and in ADMIXTURE, each line represents an individual. (a) shows the PCA and ADMIXTURE results for clustering of Roma and HapMap populations. The populations codes are as follows: Yoruba in Ibadan, Nigeria (YRI), Luhya in Webuye, Kenya (LWK), Maasai in Kinyawa, Kenya (MKK), Utah residents with Northern and Western European ancestry (CEU), Toscani in Italia (TSI), Han Chinese in Beijing, China (CHB), Japanese in Tokyo, Japan (JPT), Chinese in Metropolitan Denver, Colorado (CHD), Gujarati Indians in Houston, Texas (GIH), African ancestry in Southwest USA (ASW) and Mexican ancestry in Los Angeles, California (MEX), and (b) shows the PCA and ADMIXTURE results for clustering of Roma and South Asian groups. We limit the sample size of all groups (except Roma) to 20 individuals.
Figure 2.
We performed ROLLOFF (using R(d)) on the Roma samples (n = 24). We plot the weighted covariance as a function of genetic distance, and obtain a date by fitting an exponential function with an affine term: , where d is the genetic distance in Morgans and n is the number of generations since mixture. We do not show inter-SNP intervals of <0.5 cM since we have found that at this distance admixture LD begins to be confounded by background LD.
Figure 3.
The European and South Asian sources of Roma ancestry.
We computed a genome-wide average IBD sharing distance between Roma (all samples combined in one group) and other regional groups. Details of the regional grouping are described in Methods. (a) shows the average pairwise IBD sharing between Roma and Europeans (grouped into four regional categories), (b) shows IBD sharing average pairwise IBD sharing between Roma and South Asians (grouped into 8 regional categories).
Figure 4.
(a) shows estimates of genomewide autozygosity in Roma and individuals from HapMap (n = 30 from each of the 11 HapMap populations). Each point represents an individual with the color-coding described in the legend. (b) shows the decay of autocorrelation with genetic distance. We fit an exponential function: where D = distance in Morgans and t = time of founder event. We thus infer a founder event date of 27 generations.