Human Population Genetic Structures Revealed By High Density Autosomal SNPs

The genetic structures of 40 human populations distributed in four main continents(Africa,European,Asian and Ameica)and Pacific islands were revealed by high density autosomal single nucleotide polymorphisms(SNPs). Geographical distribution of human populations was found contribute much to the observed genetic structures.Asian populations,African populations and European Populations formed three clusters in phylogenetic tree reconstructed by autosomal SNPs.Most of populations within each continent also clustered together according to their geographical distributions.For populations in the three main continents,most of genetic differences were due to variants within populations(85.6%),the difference among populations within continent was only 1.2%,which,however,was still less than the difference of populations between continents(13.2%).The distribution of genetic difference in the populations of six main language families in Eastern and Southeast Asia had the similar pattern to that of among continents. except that the difference between populations within language families (3.22%)was larger than that of between language families(1.02%),but the main difference was within populations.The language family structures were supported by genetic evidence to a large degree,but they were not always consistent.The language family's attachment of Zhuang and Miao in China has been argued for along time.It was shown they had close relationship with Han Chinese in genetics.The results of AMOVA in this study indicated that the difference between Han and Zhuang or Han and Miao was still larger than within populations.The results of AMOVA also showed the genetic relationship between Han and Titetan populations were very close,the difference between the two populations was less than 0.56%.The results of STRUCTURE analysis of 15 Asian populations suggested there are at least 9 clusters or sub-structures.Zhuang,Miao,Wa,Jinuo,Uygur and Han in China all had different genetic structures,this was also true in four populations, Bidayuh,Temuan,Jehai,Kensiu in Malaysia.However,Korean,Japanese and Ryukyu had the similar genetic structures and could not be distinguished easily.In fact,Han Chinese should an admixture population,with the main genetic components same as Korean,Japanese et al.,but it had also 34.4% genetic components same as that of Zhuang and 12%same as Miao.With the results derived from analysis of linkage disequilibrium in 11 populations,from the point of view of "LD block",differences of the LD patterns in populations reflected different histories of populations.There were large diversities of LD pattern between African populations and non-African populations,with African populations having shorter LD blocks and non-African populations having longer LD blocks.Differences were also observed between Asian populations and European populations,the former have longer LD blocks and larger chromosomal coverage of LD blocks.Within Asian populations,LD blocks in Miao and Wa,which are two minorities in Southern China,had longer average length and maximal length than those in other populations,including Han Chinese populations and Japanese population, which distribute in Northern Asia.This result might be due to the smaller effective population size of Miao and Wa.Allele frequencies of SNPs affected the results of LD block partition,although there were no substantial changes of overall pattern.The two admixture populations,African-American and Uigur had smaller LD blocks than their possible parental populations.Notably,LD blocks of African-American population were shorter than those of African population and European populations;Uigur had shorter LD blocks than Asian populations and European populations.In the term of association studies,linkage disequilibrium(LD)in admixture populations and isolated populations are very attractive and has practical meanings.Therefore,the structures of two typical admixture populations,African-American and Uigur in China,one typical isolation population,Samoan were studied here.African American population had higher heterozygosity than that of general populations,with average observed heterozygosity 0.311 in AfA, 0.296 in YRI and 0.290 in CEU;this is one of characteristic of admixture populations.We examined the extent and magnitude of LD in a sample of 48 African Americans(AfA)with those in its parental populations,i.e.60 European individuals(CEU)and 60 African individuals(YRI).We showed that the LD in AfA is similar to that in YRI but less than that In CEU using unselected markers.Furthermore,the elevation of LD in AfA,compared with its parental populations,can only be observed at the markers with high allele frequency difference between two parental populations(f≥0.4,known as ancestry informative markers,AIMs),but not at the other markers.In AfA,the AIMs contribute mainly to LD(measured by r~2)at level of 0.1≤r~2<0.8 and beyond 200kb.High level LD(r~2≥0.8)extends no more than 200kb in all three populations.Using the AIMs,we inferred the ancestral origins of chromosomal segments in AfA individuals.We further showed that those AfA individuals with ancestry dominantly from African population contribute little to admixture LD and removal of such individuals led to an increase of LD by 1.75-fold.Therefore,we proposed that the extent and magnitude of LD can be enhanced by selecting AIMs with f≥0.4 and by removing individuals derived from single ancestry in mapping genes in admixed populations.Uigur population of Hetian at Xinjiang in China is an admixture population,with both obvious European and Asian ancestries,with 52.1% European ancestries and 47.9%Asian ancestries.Therefore,Uigur is more close to European population in Genetics.Uigur had average observed heterozygosity 0.309,which was higher than those in European and Asian populations,which had 0.301 and 0.284 respectively.Admixture proportions of Uigur individuals varied much less than admixture proportions of African American populations,with the minimal 44.4%,and the maximal 61.8%.We estimated the admixture event happened 92 generations or 1840 years ago, suppose the modern Uigur population was formed by one admixture event in history.However,this estimated time is much earlier than that estimate from other evidence.Linkage disequilibrium of Uigur took no advantage over that of its parental populations,especially at level of r~2≥0.5,even for selected AIMs. This result was totally different with that observed in African American.But high level LD(r~2≥0.8)extended no more than 200kb in all three populations, i.e Uigur and its two parental populations.Contrast to the conditions in African Americans and Uigur,Samoan population has lower average observed heterozygosity than that of general populations;this is one of characteristic of isolated populations.Although long LD is attractive,unlike in admixture populations such as African-American,the interested region or distance we should focus our mind on is within 200kb, where LD in isolated populations such as Samoan elevates comparing to general populations.Allele frequencies affect observation of LD,at least when measured by r~2,i.e elevation of Samoan LD was more pronounced in common alleles(MAF≥0.15).High level LD(r~2≥0.8)consistently elevated (average 2.66-fold of CEU LD and 2.33-fold of CHB LD for common alleles)at all distance scale within 200kb,while lower level LD(1/3≤r~2＜0.5)elevated only beyond 30kb,and still lower LD(0.1≤r~2＜1/3)elevated only beyond 50kb, however,they did extend to longer distance than that in general populations.