Improved inference of site-specific positive selection under a generalized parametric codon model when there are multinucleotide mutations and multiple nonsynonymous rates

Table 3 False positive rates (null scenarios) and true positive rates (alternative scenarios) for three LRTs when the evolutionary process includes both ω variability among sites and MNRs

LRT-1 compares M1a to M2a (under-fit models). LRT-2 compares G1a^x to G2a^x (perfect-fit models). LRT-3 compares G1a¹³ to G2a¹³ (over-fit models). The asterisk symbol (*) indicates scenarios where either convergence problems or suboptimal peaks were encountered for the models of LRT-3. To overcome these problems, models were re-fit to the same dataset multiple times, each using a different set of initial parameter values. The number of problematic datasets for SNR was 2a = 21 and 2b = 1; for low MNR was 2a = 27, 2b = 16, 2c = 16 and 2f = 10; and for high MNR was 2a = 29, 2b = 20, 2c = 35, 2e = 15, 2f = 15 and 2 g = 1. Because using multiple initials for the problematic datasets was successful, the results above are for all 100 replicates

ISSN: 2730-7182