Estimation of population genetic parameters in human isolates from Vale do Ribeira, São Paulo ("Quilombo" populations)

Ca. four million Africans were brought as slaves to di erent Brazilian regions during about 300 years in colonial \ud times. Many quilombo populations were founded by runaway or abandoned African slaves in Brazilian isolated \ud areas before the end of slavery in 1888. In the state of São Paulo, most of the quilombo remnants are located in \ud its southern part, in the Vale do Ribeira region. By means of 30 autosomal molecular markers (one ALU insertion, \ud 13 SNPs and 16 microsatellites) and genealogical data analyses, we estimated population genetic parameters to \ud characterize both the breeding system as well as the degree of genetic substructure of 10 di erent communities in the \ud region. Genealogical analysis of data collected from about 2000 individuals enabled us to obtain a mean inbreeding \ud coe cient value F = 0.00134; this value is almost 1.5 times higher than the corresponding one estimated for the \ud total Brazilian population and almost twice as high the same parameter for the state of São Paulo. However, this \ud represents an underestimate of the true value, due to lack of reliable genealogical information. By means of genomic \ud loci data analysis from about 700 individuals, we obtained mean F values for the quilombo communities separately \ud and as a whole. For the communities separately, twelve (5%) out of a total of 239 tests from eight communities \ud showed that genotypes deviated from Hardy-Weiberg ratios, as expected by chance; and for the set of all quilombo \ud communities, four (13.33%) out of 30 loci deviated signi cantly from Hardy-Weinberg ratios, indicating an excess \ud of homozygosis. We also estimated a weighted mean value of F = 0,034 for the whole population by averaging the \ud F values obtained from each locus by the reciprocal of their corresponding variances. is F value is comparable \ud to that of other isolates reported in the literature. No signi cant levels of population substructure were detected, \ud since the estimated FST values (inbreeding component due to random drift within subpopulations) were in \ud general quite modest, which could be explained by signi cant levels of gene ow among quilombo populations.FAPESP and CNPq