ProteinEvolverABC2: Coestimation of recombination and substitution rates in protein sequences by approximate Bayesian computation and structural constraints

Data of the study: ProteinEvolverABC2: Coestimation of recombination and substitution rates in protein sequences by approximate Bayesian computation and structural constraints.

Quantifying mutation and recombination is essential for understanding the strength of evolutionary processes. While these parameters are commonly estimated from nucleotide sequences, their estimation from protein sequences has been less explored, especially for recombination, despite the widespread use of protein sequencing. ProteinEvolverABC2 co-estimates substitution and recombination rates from protein sequences using the approximate Bayesian computation approach, with and without regression adjustments. The framework incorporates a variety of fixed and nuisance population genetics parameters, as well as evolutionary constraints derived from the protein structure, thereby extending its previous version. It showed accurate co-estimation of both parameters at typical levels under different substitution models of protein evolution. As expected, performance decreases at extreme levels of sequence diversity due to reduced genetic signatures of substitution and recombination events and, in such scenarios, accounting for structural evolutionary constraints can partially improve the estimates. As illustrative examples, we applied the framework to viral protein families of general interest.

This work was supported by the Grants PID2023-151032NB-C22 and PID2023-151032NB-C21 funded by MCIU/AEI/10.13039/501100011033 and by FEDER, UE.