The impact of non-neutral synonymous mutations when inferring selection on non-synonymous mutations

Area of Science:

• Evolutionary genetics
• Population genetics
• Molecular evolution

Background:

• The distribution of fitness effects (DFE) quantifies the impact of new mutations on reproductive fitness, a key parameter in evolutionary genetics.
• Current methods infer the DFE for nonsynonymous mutations by first estimating demographic parameters from synonymous variants, assuming neutrality.
• Emerging evidence suggests synonymous mutations may also be subject to selection, potentially violating this neutrality assumption.

Purpose of the Study:

• To investigate whether selection acting on synonymous mutations biases the inference of the DFE for nonsynonymous mutations.
• To assess the impact of unmodeled selection on synonymous variants on demographic parameter estimation and subsequent DFE inference.

Main Methods:

• Simulated various models of selection on synonymous mutations using SLiM.
• Inferred the DFE of nonsynonymous mutations using Fit∂a∂i, a standard computational method.
• Compared DFE and demographic parameter inferences under different selection scenarios for synonymous variants.

Main Results:

• Selection on synonymous variants led to inaccurate inferences of recent population growth.
• Under specific conditions, this bias resulted in an inflated proportion of highly deleterious nonsynonymous mutations in DFE estimates.
• The detrimental effects on DFE inference were mitigated when correct demographic parameters were utilized instead of those derived from selected synonymous variants.

Conclusions:

• Unaccounted selection on synonymous mutations can significantly distort downstream inferences of the DFE for nonsynonymous mutations.
• Accurate demographic parameter estimation is crucial for reliable DFE inference, highlighting the need to consider selection on synonymous sites.
• This study underscores the importance of modeling selection on synonymous variants to avoid biases in fundamental evolutionary genetic parameters.