On the causes of evolutionary transition:transversion bias

Arlin Stoltzfus; Ryan W. Norris

Most predictors are more powerful than the ti:tv distinction (AUC = 0.53). The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/027722 doi: bioRxiv preprint Figure 3 . Relationship between power to predict mutant fitnesses and evolutionary effect size. AUC and evolutionary acceptance ratio are shown for increasingly randomized verions of U B . For the unrandomized U B , the power in predicting mutational effects is AUC = 0.64, and this corresponds to an evolutionary acceptance ratio of 2.7 for conservative versus radical replacements. To estimate the evolutionary acceptance ratio for more modest values of AUC, we can weaken U B by randomly re-assigning "conservative" or "radical" labels to an increasingly large fraction of replacement types The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/027722 doi: bioRxiv preprint Figure 4 . The relative advantage of transitions as indicated by a random sample of biochemical factors. Each biochemical attribute of an amino acid is converted to a pairwise similarity measure, so that each possible amino acid replacement has a similarity score. Here the AUC is the chance that a replacement due to a randomly chosen transition (from the pool of actual mutants from the 8 studies) has a higher similarity score (for the given biochemical attribute) than a randomly chosen transversion. The resulting distribution indicates that many biochemical factors make transitions seem more conservative (AUC > 0.5), but a roughly equal number make transversions seem more conservative (AUC < 0.5). For the full set of predictors, see 24 0.24 0.017 * 0.37 1.7e-03 ** 0.51 9.4e-05 ** Combined 1239 0.047 6.3e-07 ** 0.088 9.4e-21 ** 0.090 3.2e-16 ** (a) P values for the null hypothesis of no correlation (*, P < 0.05; **, P < 0.01).