However, significant increases in GC content were observed in the first and second codon positions as well.
The GC content of the first and second codon positions increased as well, but to a lesser extent.
The sea squirt genes showed lower GC1, higher GC2 and lower GC3 (GC1 and GC2 denote G+C content at the first and second codon positions respectively) compared to fish orthologs (Table 1).
Considering only the b-like alleles of S. invicta, a greater than 3-fold excess of coding-region substitutions exists, attributable mostly to augmented variation in the first and second codon positions.
Since nucleotide changes in the first and second codon positions usually result in changes in the amino acid sequence, these results suggested that genomic GC content has a significant impact protein amino acid sequence.
Since genomes with high GC content have increased GC content in the first and second codon positions, there is increased use of the three GC-rich codon families and decreased use of the five GC-poor codon families in protein coding regions.
Evolutionary model parameters incorporated in BEAST included a SDR06 model of nucleotide substitution, which uses a HKY substitution model with a Gamma site heterogeneity and two codon partitions (one for the first and second codon positions and a second for the third).
By estimating the node sequences at the nucleotide level, we calculated the numbers of nonsynonymous changes at the first and second codon sites, and also the number of synonymous changes at the FD3rd codon sites, between the root sequence and the terminal sequences.
Moreover, analysis by codon position revealed that this variation in G+C richness was entirely confined to the third codon position (Figure 2B; PTB 86%, ROD1, 35%, nPTB 32%), whereas the first and second codon positions had nearly identical nucleotide composition (not shown).
Therefore, only the first and second codon positions were included in our analysis.
