Fig. 5
Folding bias (∆LFE) is positively correlated with genomic GC-content throughout the CDS. a The effect of genomic-GC on ∆LFE at each position along the CDS start (left) and end (right), measured using GLS regression R2 values. R2 values above the X-axis indicate positive regression slope (indicating moderating effect of GC-content); R2 values below the X-axis indicate negative regression slope (i.e., reinforcing effect of GC-content). Near the CDS edges (where ∆LFE is usually positive), genomic-GC generally has a moderating effect on ∆LFE. In the mid-CDS region (where ∆LFE is usually negative), genomic-GC generally has a reinforcing effect on ∆LFE. Major taxonomic groups are plotted as different colored lines. White dots indicate regression p value < 0.01. b Comparison of ∆LFE profile values in species with high vs. low genomic GC-content. Species with high GC-content (blue, genomic-GC > 45%) tend to have more extreme ∆LFE and show the conserved ∆LFE regions more clearly, while species with low GC-content (yellow, genomic-GC ≤ 45%) tend to also have weak ∆LFE. c. Genomic GC-content for all species plotted on the PCA coordinates of their ∆LFE profiles (same coordinates as in Fig. 3b. N = 513) for CDS start (left) and end (right). Low-GC species are generally clustered in a small region, indicating they have similar ∆LFE profiles, and that region is characterized by weak ∆LFE. d Qualitative summary of ∆LFE in relation to GC-content in the mid-CDS