Population stratification analysis based on ancestry-informative single-nucleotide polymorphisms (AISNPs) is a fundamental approach in forensic biogeographic inference. Although numerous markers have been developed to distinguish major continental populations, accurate inference at the subcontinental or intraregional level remains challenging and generally requires high-density SNP datasets generated by massively parallel sequencing or microarray platforms.
In the present study, we assessed the feasibility of discriminating among three East Asian populations—Japanese, Korean, and Chinese—using the Axiom Japonica Array NEO (Thermo Fisher Scientific), which comprises approximately 650,000 SNPs. Reference genotype data were obtained from the 1000 Genomes Project (104 Japanese [JPT], 103 Han Chinese in Beijing [CHB], and 163 Southern Han Chinese [CHS]) and the Korea Genomics Center (KOGIC) (3,378 Korean individuals).
From each population, 100 individuals were randomly selected. Of these, 80 individuals per group were used for marker selection, and the remaining 20 were reserved for validation. SNPs with a minor allele frequency below 0.01 were excluded. Subsequent analysis of variance (ANOVA) with Bonferroni correction identified a panel of 2,500 SNPs from the original ~650,000 SNPs that were sufficient to effectively discriminate the three populations. Validation analysis demonstrated accurate classification of individuals into their respective populations.
Given that forensic specimens are often highly degraded and limited in quantity, we further evaluated the sensitivity of the microarray system. Serial dilutions of genomic DNA (100 ng to 1 ng) were analyzed, revealing that as little as 1 ng of input DNA was sufficient to successfully genotype approximately 600,000 SNPs.
These findings indicate that a reduced set of 2,500 SNPs can serve as an effective AISNP panel for the discrimination of closely related East Asian populations, providing a robust and sensitive tool applicable even to trace, degraded forensic DNA samples.