Inference of Population Structure using Dense Haplotype Data

摘要: The advent of genome-wide dense variation data provides an opportunity to investigate ancestry in unprecedented detail, but presents new statistical challenges. We propose a novel inference framework that aims efficiently capture information on population structure provided by patterns haplotype similarity. Each individual sample is considered turn as recipient, whose chromosomes are reconstructed using chunks DNA donated the other individuals. Results this “chromosome painting” can be summarized “coancestry matrix,” which directly reveals key about ancestral relationships among If markers viewed independent, we show matrix almost completely captures used both standard Principal Components Analysis (PCA) and model-based approaches such STRUCTURE unified manner. Furthermore, when linkage disequilibrium, combines across successive increase ability discern fine-scale PCA. In parallel, have developed efficient approach identify discrete populations matrix, offers advantages over PCA terms interpretability existing clustering algorithms speed, number separable populations, sensitivity subtle structure. analyse Human Genome Diversity Panel for 938 individuals 641,000 markers, 226 reflecting differences continental, regional, local, family scales. present multiple lines evidence that, while many methods similar strongly differentiated groups, more human consistently at much finer level than currently available geographic labels only captured haplotype-based approach. software article, ChromoPainter fineSTRUCTURE, from http://www.paintmychromosomes.com/.