Figure 5.

The patterns of the correlation slope (A, C, and E) and the coefficient of determination R² values (B, D, and F) for codon usages between different GC groups. Fifty-codon (A and B), CDS (C and D), and whole genome (E and F) were used respectively as an analysis unit, with a GC-content difference less than 0.25 (the values are nearly 0 when ΔGC > 0.25). As shown in 50 consecutive codons (A and B) and CDS (C and D) analysis units, when the difference of GC-content from two compared units is 0 (x-axis), both the linear regression slope and the regression coefficient R² are infinitely approaching 1 (P = 2.32e-54), which strongly suggests that codon usage frequencies are nearly constant in regions with similar GC-content.