Thursday, October 27, 2016

BISTRO - Zhengting Zou

4:00 PM

2036 Palmer Commons

BISTRO is restricted to U-M Bioinformatics Graduate Program students and faculty.

"Comparative analysis of non-synonymous transition:transversion bias and amino acid exchangeability across the tree-of-life."

Abstract

Transition:transversion bias in nucleotide sequence evolution has been observed in numerous species. Despite a mutational bias that transition is usually easier to happen, whether transition is also favored by selection in coding sequences is still unclear. In this study, we modified a previous codon evolution model to address potential selectional bias (η) towards transition and implemented a maximum likelihood estimation of this bias parameter η. (1) We examined genome-wide coding sequence alignments from a wide variety of sister species clades and found that different species clades on the tree of life have different η, i.e., selectional bias can favor either transitions or transversions depending on the species under study. (2) We conducted multiple simulations to show that many parameters in the model can affect the bias η. Surprisingly, the observed variation of η seems to be mainly contributed by the fitness effects of amino acid changes, which people usually assume to be constant across different species. (3) We further infer these fitness effects in multiple species and show different species have different fitness effects for the same amino acid change.