PMID- 33586030 OWN - NLM STAT- MEDLINE DCOM- 20210827 LR - 20211204 IS - 1437-1596 (Electronic) IS - 0937-9827 (Linking) VI - 135 IP - 4 DP - 2021 Jul TI - Population inference based on mitochondrial DNA control region data by the nearest neighbors algorithm. PG - 1191-1199 LID - 10.1007/s00414-021-02520-3 [doi] AB - Population and geographic assignment are frequently undertaken using DNA sequences on the mitochondrial genome. Assignment to broad continental populations is common, although finer resolution to subpopulations can be less accurate due to shared genetic ancestry at a local level and members of different ancestral subpopulations cohabiting the same geographic area. This study reports on the accuracy of population and subpopulation assignment by using the sequence data obtained from the 3070 mitochondrial genomes and applying the K-nearest neighbors (KNN) algorithm. These data also included training samples used for continental and population assignment comprised of 1105 Europeans (including Austria, France, Germany, Spain, and England and Caucasian countries), 374 Africans (including North and East Africa and non-specific area (Pan-Africa)), and 1591 Asians (including Japan, Philippines, and Taiwan). Subpopulations included in this study were 1153 mitochondrial DNA (mtDNA) control region sequences from 12 subpopulations in Taiwan (including Han, Hakka, Ami, Atayal, Bunun, Paiwan, Puyuma, Rukai, Saisiyat, Tsou, Tao, and Pingpu). Additionally, control region sequence data from a further 50 samples, obtained from the Sigma Company, were included after they were amplified and sequenced. These additional 50 samples acted as the "testing samples" to verify the accuracy of the population. In this study, based on genetic distances as genetic metric, we used the KNN algorithm and the K-weighted-nearest neighbors (KWNN) algorithm weighted by genetic distance to classify individuals into continental populations, and subpopulations within the same continent. Accuracy results of ethnic inferences at the level of continental populations and of subpopulations among KNN and KWNN algorithms were obtained. The training sample set achieved an overall accuracy of 99 to 82% for assignment to their continental populations with K values from 1 to 101. Population assignment for subpopulations with K assignments from 1 to 5 reached an accuracy of 77 to 54%. Four out of 12 Taiwanese populations returned an accuracy of assignment of over 60%, Ami (66%), Atayal (67%), Saisiyat (66%), and Tao (80%). For the testing sample set, results of ethnic prediction for continental populations with recommended K values as 5, 10, and 35, based on results of the training sample set, achieved overall an accuracy of 100 to 94%. This study provided an accurate method in population assignment for not only continental populations but also subpopulations, which can be useful in forensic and anthropological studies. FAU - Yang, Fu-Chi AU - Yang FC AD - Department of Forensic Medicine, College of Medicine, National Taiwan University, No.1 Jen-Ai Road Section 1, Taipei, 10051, Taiwan. FAU - Tseng, Bill AU - Tseng B AD - Department of Forensic Medicine, College of Medicine, National Taiwan University, No.1 Jen-Ai Road Section 1, Taipei, 10051, Taiwan. FAU - Lin, Chun-Yen AU - Lin CY AD - Institute of Forensic Medicine, Ministry of Justice, New Taipei City, 23016, Taiwan. FAU - Yu, Yu-Jen AU - Yu YJ AD - Department of Forensic Medicine, College of Medicine, National Taiwan University, No.1 Jen-Ai Road Section 1, Taipei, 10051, Taiwan. FAU - Linacre, Adrian AU - Linacre A AD - College of Science & Engineering, Flinders University, Adelaide, 5001, Australia. FAU - Lee, James Chun-I AU - Lee JC AD - Department of Forensic Medicine, College of Medicine, National Taiwan University, No.1 Jen-Ai Road Section 1, Taipei, 10051, Taiwan. jimlee@ntu.edu.tw. LA - eng GR - MOST 107-2320-B-002-045-MY3/Ministry of Science and Technology, Taiwan/ PT - Journal Article DEP - 20210214 PL - Germany TA - Int J Legal Med JT - International journal of legal medicine JID - 9101456 RN - 0 (DNA, Mitochondrial) SB - IM MH - *Algorithms MH - DNA, Mitochondrial/*genetics MH - Genetics, Population/*methods MH - Humans MH - Indigenous Peoples/genetics MH - *Locus Control Region MH - *Phylogeny MH - Racial Groups/*genetics MH - Taiwan/ethnology OTO - NOTNLM OT - Control region OT - Nearest neighbors algorithm OT - Population inference OT - mtDNA EDAT- 2021/02/16 06:00 MHDA- 2021/08/28 06:00 CRDT- 2021/02/15 06:14 PHST- 2020/11/30 00:00 [received] PHST- 2021/01/27 00:00 [accepted] PHST- 2021/02/16 06:00 [pubmed] PHST- 2021/08/28 06:00 [medline] PHST- 2021/02/15 06:14 [entrez] AID - 10.1007/s00414-021-02520-3 [pii] AID - 10.1007/s00414-021-02520-3 [doi] PST - ppublish SO - Int J Legal Med. 2021 Jul;135(4):1191-1199. doi: 10.1007/s00414-021-02520-3. Epub 2021 Feb 14.