Differentiation of Asian population samples using the Illumina ForenSeq kit

• Published in: Forensic science international : genetics 2020-09, Vol.48, p.102318-102318, Article 102318
• Main Authors: Ramani, Anantharaman, Shue, Bing Hong, Tan, Si Zhen, Wong, Yongxun, Syn, Christopher
• Format: Article
• Language: English
• Subjects: Asian populations; Bio-geographic ancestry prediction; ForenSeq; Singapore; Snipper; STRUCTURE
• ISSN: 1872-4973; 1878-0326
• DOI: 10.1016/j.fsigen.2020.102318

•Ancestry prediction of 1,234 Asian samples from 8 populations in Singapore.•48 ancestry SNPs and 12 phenotypic SNPs with AIM properties shortlisted.•Ancestry modelling using STRUCTURE shows 8 populations are better grouped as 5.•Poor self-classification (69 %) using Snipper.•60 AIMs less effective of differentiation of Asian populations beyond East, South and South-East Asian population groups. The Kidd set of ancestry informative SNPs are included in Illumina’s ForenSeq DNA Signature Kit. We had previously reported on the capability of these SNPs together with some phenotypic SNPs with ancestry informative properties in differentiating individuals from the Chinese, Malay and Indian populations in Singapore. The Singapore population is primarily made up of Chinese, Malays and Indians, with individuals from other Asian countries making up the rest. In this study, we evaluated the ancestry prediction capabilities of the ForenSeq kit in 484 unrelated individuals of self-declared Bangladeshi, Burmese, Filipino, Indonesian and Vietnamese origin. 750 Chinese, Malay and Indian individuals previously reported were included in this study. 48 ancestry SNPs and 12 phenotypic SNPs with ancestry informative properties were selected for analyses. Ancestry modelling in STRUCTURE showed that the eight tested populations could be better classified as five. Principal component analysis also showed that the eight populations clustered in five groups based on general geographic location within Asia; with Chinese clustering with Vietnamese, Malays clustering with Indonesians, Indians clustering with Bangladeshi, and the Burmese and Filipino populations clustering in-between and overlapping with the Chinese and Malay populations. The 60 SNPs analysed could account for only 23 % of the variation between the populations. The lack of distinction between the populations resulted in poor (43 % correct self-classification) cross-validation using Snipper. While this was improved by merging the co-clustering populations into five groups (East, South-East, South Asian, Burmese & Filipino), successful self-classification was still relatively low (69 %). While the 60 tested ancestry informative markers were able to differentiate between individuals of East, South-East and South Asian origin, they are not sufficiently informative to effectively discriminate between Chinese, Malays and Indians, and Bangladeshi, Burmese, Filipino, Indonesian and Filipino populations in the country.