Figure 6.

Effect of AT content of training set upon translation accuracy. Each purple diamond represents a complete CDS set from a prokaryote genome. The orange box represents all CDS available from the nematode order Spirurida (~230000 non-redundant coding nucleotides). The green triangle represents the complete Arabidopsis thaliana RefSeq collection (~30000000 non-redundant coding nucleotides). The green circles are training sets of A. thaliana CDS RefSeq entries randomly selected to total ~230000 non-redundant coding nucleotides. The AT content of C. elegans is shown by the vertical dashed line.