Isolating selective from non-selective forces using site frequency ratios.

Area of Science:

• Population Genetics
• Evolutionary Biology
• Bioinformatics

Background:

• Estimating mutation fitness effects is crucial for understanding evolution.
• Previous methods often rely on assumptions about non-selective factors or incorporate them into complex models.
• Accurate estimation of selection effects is challenged by demographic factors and linked selection.

Purpose of the Study:

• To introduce a novel method (SFRatios) for estimating the distribution of mutation fitness effects.
• To overcome limitations of existing methods by minimizing assumptions about non-selective factors.
• To improve the accuracy of statistical tests for selection and fitness effect estimation.

Main Methods:

• Developed the SFRatios method using ratios of selected to neutral sites to mitigate non-selective effects.
• Derived a likelihood expression by approximating the ratio of two Poisson random variables as a ratio of Gaussian random variables.
• Evaluated performance using simulations across diverse demographic models, including linked selection effects.

Main Results:

• The SFRatios method demonstrates good performance in statistical tests of selection and estimation of selection effects.
• Performance is particularly strong for weak selection, expansion, and structured demographic models compared to bottleneck models.
• Application to Drosophila melanogaster populations revealed weak selection on synonymous sites and consistent, stronger selection on nonsynonymous sites in an African population.

Conclusions:

• The SFRatios method provides a robust approach for estimating mutation fitness effects, reducing reliance on assumptions about non-selective factors.
• The method is effective across a range of demographic scenarios and offers insights into selection acting on different site types.
• Findings in Drosophila highlight subtle but significant selective pressures, varying by population and site type.