Human Genetics

, Volume 117, Issue 6, pp 511–519

Use of autosomal loci for clustering individuals and populations of East Asian origin


  • Jong-Jin Kim
    • National Institute of Scientific InvestigationDNA Analysis Division
  • Paul Verdu
    • Department of GeneticsYale University School of Medicine
  • Andrew J. Pakstis
    • Department of GeneticsYale University School of Medicine
  • William C. Speed
    • Department of GeneticsYale University School of Medicine
  • Judith R. Kidd
    • Department of GeneticsYale University School of Medicine
    • Department of GeneticsYale University School of Medicine
Original Investigation

DOI: 10.1007/s00439-005-1334-8

Cite this article as:
Kim, J., Verdu, P., Pakstis, A.J. et al. Hum Genet (2005) 117: 511. doi:10.1007/s00439-005-1334-8


We studied the genetic relationships among East Asian populations based on allele frequency differences to clarify the relative similarities of East Asian populations with a specific focus on the relationships among the Koreans, the Japanese, and the Chinese populations known to be genetically similar. The goal is to find markers appropriate for differentiating among the specific populations. In this study, no prior data existed for Koreans and the markers were selected to differentiate Chinese and Japanese. We typed, using AB TaqMan assays, single-nucleotide polymorphisms (SNPs) at 43 highly selected mostly independent diallelic sites, on 386 individuals from eight East Asian populations (Han Chinese from San Francisco, Han Chinese from Taiwan, Hakka, Koreans, Japanese, Ami, Atayal, and Cambodians) and one Siberian population (Yakut). We inferred group membership of individuals using a model-based clustering method implemented by the STRUCTURE program and population clustering by using computer programs DISTANCE, NEIGHBOR, LSSEARCH, and DRAWTREE, respectively, calculating genetic distances among populations, calculating neighbor-joining and least-squares trees, and drawing the calculated trees. On average 52% of individuals in the three Chinese groups were assigned into one cluster, and, respectively, 78 and 69% of Koreans and Japanese into a different cluster. Koreans differentiated from the Chinese groups and clustered with the Japanese in the principal component analysis (PCA) and in the best least-squares tree. The majority of Koreans were difficult to distinguish from the Japanese. This study shows that a relatively few highly selected markers can, within limits, differentiate between closely related populations.

Supplementary material

439_2005_1334_ESM_supp.pdf (371 kb)
(PDF 372 KB)

Copyright information

© Springer-Verlag 2005