A composite-likelihood approach for detecting directional selection from DNA sequence data.

摘要: We present a novel composite-likelihood-ratio test (CLRT) for detecting genes and genomic regions that are subject to recurrent natural selection (either positive or negative). The method uses the likelihood functions of Hartl et al. (1994) inference in Wright-Fisher genic model corrects nonindependence among sites by application coalescent simulations with recombination. Here, we (1) characterize distribution CLRT statistic (Λ) as function population recombination rate (R = 4Ner); (2) explore effects bias estimation R on size (type I error) CLRT; (3) robustness growth, bottlenecks, migration; (4) power under varying levels mutation, selection, recombination; (5) discriminatory distinguishing negative from growth; (6) evaluate performance maximum composite-likelihood (MCLE) coefficient. find has excellent detect weak moderate selection. Moreover, is quite robust estimate local rate, but not certain demographic scenarios such growth recent bottleneck. Last, demonstrate MCLE parameter little downward positively selected mutations.