Estimating the selection pressure and evolutionary rate of proteins on the non-neutral hypothesis of synonymous mutations

The Nonsynonymous/Synonymous substitution rate ratio (Ka/Ks) is a commonly used metric to estimate the selection pressure and evolutionary rate of proteins in comparative genomics, which plays critical roles in molecular evolution in both biology and medicine. A fundamental assumption of Ka/Ks is that synonymous mutations are evolutionarily neutral and not subject to natural selection as they do not alter protein sequences and function. However, a number of studies have demonstrated that synonymous mutations are non-neutral and may lead to diseases through a number of mechanisms, such as altering miRNA regulation. This further implies that synonymous mutations also participate in the process of natural selection and thus Ka/Ks should be redefined as well. For this purpose, here we propose an improved Ka/Ks ratio, i Ka/Ks, which re-computed the neutral substitution rate by taking the altered status of miRNA binding into consideration, and thereby incorporate the impact of synonymous mutations on miRNA regulation. As a result, i Ka/Ks shows better performance than Ka/Ks when comparing them using their correlation with expression distance. Moreover, case studies showed that i Ka/Ks is able to identify the positive/negative selection genes that are missed by Ka/Ks. For example, TMEM72/Tmem72 is estimated to be positively selected by i Ka/Ks (1.13) but negatively selected by the conventional Ka/Ks ratio (0.21). Further evidence showed its rapid evolution, which further support the power of the new algorithm.