In this study, researchers explore large-scale selection of highly informative microhaplotypes for ancestry inference and population specific informativeness.
This study presents a comprehensive way of selecting highly informative microhaplotypes (MHs) for accurate ancestry inference through the development of a pipeline to efficiently select such MHs from large-scale genomic datasets. MHs describe physically close genetic markers that are inherited together and are gaining prominence due to their efficiency in forensic, clinical, and population studies. They excel in kinship analysis, DNA mixture detection, and ancestry inference, offering advantages in precision over individual SNPs and STRs. Over 120,000 MHs were identified from almost a million markers, which allow this non-independent information to be efficiently used for inference. The MHs were compared to SNPs in terms of their informativeness and performance of their subsets in ancestry inference and all the results consistently favored MHs. A method for ranking markers by specific population informativeness was also introduced, which showed improvement in the accuracy of Native American ancestry estimation, overcoming the challenges of its underrepresentation in datasets. The proposed approach and the subsets selected by specific population informativeness offer valuable tools for improving ancestry inference accuracy, particularly for admixed populations as demonstrated for a Brazilian dataset. (Published Abstract Provided)