Skip to main content
Fig. 1 | BMC Evolutionary Biology

Fig. 1

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

Fig. 1

Graphical illustration of the design of Simulation Studies 1 and 2. The overall design is comprised of 32 distinct evolutionary scenarios divided into four distinct Simulation studies focused on different objectives. The details of Simulation studies 1 and 2 are shown in this figure. The details of Simulation Studies 3 and 4 are derived from Studies 1 and 2, and are further explained in the text. All simulation studies were comprised of 100 replicates, each having sequences of 300 codons. All datasets were generated using version 1.2 of the COLD program “www.mathstat.dal.ca/~tkenney/Cold/, https://github.com/tjk23/COLD”. a The 5-taxon and 17-taxon tree topologies. The 5-taxon tree and branch lengths are the same as those used for simulating sequences in Wong et al. [45]. The 17-taxon tree and branch lengths are the same as those used for simulating sequences in Yang et al. [6]. The scale for the branch lengths gives the mean number of substitutions per codon. b Sequence generating process for Simulation Study 1. The purpose of this study is to investigate the impact of DT changes (0.06 and 0.03 respectively) on the false positive rate. The selective regime is based on a strictly neutral model having just two classes of sites; conserved (50% of data) having ω = 0 and neutral (50% of data) having ω = 1. The scenarios of this study differ in the complexity of the nucleotide substitution process; case 1a is simple (everything equal) and case 1b/1c is complex (unequal GTR exchangeabilities and nucleotide frequencies). The GTR exchangeabilities and nucleotide frequencies for case 1b/1c were obtained from β-globin gene sequences. c Sequence generating process for Simulation Study 2. This study has 24 scenarios, and covers more complexity than the strictly neutral case of Study 1. Each has a mixture of three selective regimes: a large fraction of strong purifying selection (77%, ω0 = 0.05), a moderate fraction of sites (22%) with ω1 = 0.5 or 1.0, and a small fraction evolving with ω ≥ 1 (3% ω+ = 1.0, 1.5, 2.0 or 5.0). MNRs were induced using hydrophobicity factors \( \left({e}^{\beta_{HI}}\right) \) of 1.0, 0.4 or 0.05, which were linked to the codon model via the GPP parameter βHI. When \( {e}^{\beta_{HI}}=1 \) there is no impact on nonsynonymous rates, yielding a SNR codon model. When \( {e}^{\beta_{HI}}<1 \), codon evolution has MNRs. The nucleotide process had heterogeneous GTR exchangeabilities, and unequal nucleotide frequencies at the three positions of the codon. DT codon changes were not included in Study 2; DT was added to MNRs in Simulation Study 3

Back to article page