Sequence variations in the public human genome data reflect a bottlenecked population history

Abstract

Single-nucleotide polymorphisms (SNPs) constitute the great majority of variations in the human genome, and as heritable variable landmarks they are useful markers for disease mapping and resolving population structure. Redundant coverage in overlaps of large-insert genomic clones, sequenced as part of the Human Genome Project, comprises a quarter of the genome, and it is representative in terms of base compositional and functional sequence features. We mined these regions to produce 500,000 high-confidence SNP candidates as a uniform resource for describing nucleotide diversity and its regional variation within the genome. Distributions of marker density observed at different overlap length scales under a model of recombination and population size change show that the history of the population represented by the public genome sequence is one of collapse followed by a recent phase of mild size recovery. The inferred times of collapse and recovery are Upper Paleolithic, in agreement with archaeological evidence of the initial modern human colonization of Europe.

Maximum-Likelihood Estimation of Demographic Parameters Using the Frequency Spectrum of Unlinked Single-Nucleotide Polymorphisms

“…In scenarios of extreme growth, there is also a severe bias in f rec , even with large amounts of data. a bottlenecked history for this European data set, which is consistent with previous studies (Marth et al 2003, However, T can be estimated with more modest amounts of data, and T is not subject to the bias seen in f rec , 2004) and the "Out of Africa" model for human population history (Harpending et al 1998). Since the Seattle indicating that one may obtain reasonable estimates of the time of population size-change events, even if the SNPs European data set is composed of the same coding loci as the Seattle SNPs African-American data set, it magnitude is biased.…”

Haplotype Trees and Modern Human Origins

“…Another method to circumvent the confoundment of selection with demography is to use crossvalidation across multiple loci, just as was done above with nested-clade analysis. The mtDNA inference of a population size expansion between 30,000-130,000 years ago is cross-validated by some nuclear loci (Marth et al, 2003(Marth et al, , 2004, but others do not (Harpending and Rogers, 2000). An analysis of 10 noncoding DNA regions (to minimize selection) found no evidence for significant population size expansion in Africa, and only nominal significance (without correcting for multiple testing) in Eurasian samples (Pluzhnikov et al, 2002).…”

Nucleotide Variation and Haplotype Diversity in a 10-kb Noncoding Region in Three Continental Human Populations

“…A recent analysis of SNP density distribution in the human genome under a model including parameters of recombinaiton and population size suggested a collapse followed by a mild population expansion during the Upper Paleolithic period (Marth et al 2003). The scenario of bottleneck and expansion was further examined in Eswaran et al (2005).…”

“…Note that the greater diversity in the African sample could be caused by the larger effective population size in Africa (Nachman et al 1996). However, the smaller effective non-African population size itself might be due to the mild bottleneck event(s) in non-Africans (Marth et al 2003;Eswaran et al 2005).…”