Bayesian Clustering Using Hidden Markov Random Fields in Spatial Population Genetics

摘要: We introduce a new Bayesian clustering algorithm for studying population structure using individually geo-referenced multilocus data sets. The is based on the concept of hidden Markov random field, which models spatial dependencies at cluster membership level. argue that (i) chain Monte Carlo procedure can implement efficiently, (ii) it detect significant geographical discontinuities in allele frequencies and regulate number clusters, (iii) check whether clusters obtained without use priors are robust to hypothesis discontinuous variation frequencies, (iv) reduce loci required obtain accurate assignments. illustrate discuss implementation issues with Scandinavian brown bear human CEPH diversity panel set.