Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 Mar;139(2):531-540.
doi: 10.1007/s00414-024-03406-w. Epub 2025 Jan 8.

A machine learning approach for estimating Eastern Asian origins from massive screening of Y chromosomal short tandem repeats polymorphisms

Haeun You  1 , Soong Deok Lee  1   2 , Sohee Cho  3
Affiliations

A machine learning approach for estimating Eastern Asian origins from massive screening of Y chromosomal short tandem repeats polymorphisms

Haeun You et al. Int J Legal Med. 2025 Mar.

Abstract

Inferring the ancestral origin of DNA evidence recovered from crime scenes is crucial in forensic investigations, especially in the absence of a direct suspect match. Ancestry informative markers (AIMs) have been widely researched and commercially developed into panels targeting multiple continental regions. However, existing forensic ancestry inference panels typically group East Asian individuals into a homogenous category without further differentiation. In this study, we screened Y chromosomal short tandem repeat (Y-STR) haplotypes from 10,154 Asian individuals to explore their genetic structure and generate an ancestry inference tool through a machine learning (ML) approach. Our research identified distinct genetic separations between East Asians and their neighboring Southwest Asians, with tendencies of northern and southern differentiation observed within East Asian populations. All machine learning models developed in this study demonstrated high accuracy, with the Asian classification model achieving an optimal performance of 82.92% and the East Asian classification model reaching 84.98% accuracy. This work not only deepens the understanding of genetic substructures within Asian populations but also showcases the potential of ML in forensic ancestry inference using extensive Y-STR data. By employing computational methods to analyze intricate genetic datasets, we can enhance the resolution of ancestry in forensic contexts involving Asian populations.

Keywords: Biogeographical origin; East Asia; Machine learning; Short tandem repeat; Y chromosome.

PubMed Disclaimer

Conflict of interest statement

Declarations. Ethical approval: Approval was obtained from the Institutional Review Board of Seoul National University Hospital Biomedical Research Institute (IRB No. 1404-068-572). The procedures used in this study adhere to the tenets of the Declaration of Helsinki. Consent to participate: Informed consent was obtained from all individual participants included in the study. Competing interests: The authors have no competing interests to declare that are relevant to the content of this article.

Figures

Fig. 1
Fig. 1
Confusion matrices of the best-performing models for Asian classification: (a) OAO-SVM, (b) OAA-SVM, (c) XGBoost, and (d) Random Forest. * Darker shades represent higher counts
Fig. 2
Fig. 2
F1-score of each geographical group in Asian classification models
Fig. 3
Fig. 3
Feature importance of the best-performing random forest models for Asian classification
Fig. 4
Fig. 4
Confusion matrices of the best-performing models for East Asian classification: (a) SVM, (b) XGBoost, and (c) Random Forest. * Darker shades represent higher counts
Fig. 5
Fig. 5
F1-score of each geographical group in East Asian classification models
Fig. 6
Fig. 6
Feature importance of the best-performing random forest models for East Asian classification
See this image and copyright information in PMC

References

    1. Pan ZQ, Xu SH (2020) Population genomics of east Asian ethnic groups. Hereditas 157:49. 10.1186/s41065-020-00162-w - PMC - PubMed
    1. Li CX, Pakstis AJ, Jiang L, Wei YL, Sun QF, Wu H, Bulbul O, Wang P, Kang LL, Kidd JR, Kidd KK (2016) A panel of 74 AISNPs: Improved ancestry inference within Eastern Asia. Forensic Sci International: Genet 23:101–110. 10.1016/j.fsigen.2016.04.002 - PubMed
    1. Cao Y, Zhu Q, Huang Y, Li X, Wei Y, Wang H, Zhang J (2022) An efficient ancestry informative SNPs panel for further discriminating east Asian populations. Electrophoresis 43:1774–1783. 10.1002/elps.202100349 - PubMed
    1. Sun K, Yao Y, Yun L, Zhang C, Xie J, Qian X, Tang Q, Sun L (2022) Application of machine learning for ancestry inference using multi-InDel markers. Forensic Sci International: Genet 59:102702. 10.1016/j.fsigen.2022.102702 - PubMed
    1. Du R, Xiao C, Cavalli-Sforza L (1997) Genetic distances between Chinese populations calculated on gene frequencies of 38 loci. Sci China Ser C: Life Sci 40:613–621. 10.1007/BF02882691 - PubMed

LinkOut - more resources