Walid Mawass

Walid Mawass

he/him

Evolutionary geneticist studying how population history shapes genetic variation and trait architecture — from adaptation in wild populations to prediction in human complex-trait genetics.

Linked selection and the neutral diversity we actually observe

Long-term studies of wild populations regularly report substantial heritable variance in fitness. When fitness differences are inherited, uneven reproductive success lowers effective population size and reduces neutral diversity across the whole genome through background selection, including from unlinked sites. Each observed diversity deficit therefore implies a certain amount of selected fitness variance.

Published data from 19 monitored vertebrate populations do not fit that expectation in either direction. The reported fitness variances are too small to explain the observed deficits, and at the same time too large to be maintained under mutation–selection–drift balance. Measurement bias cannot explain the paradox: adjusting the estimates in either direction closes one gap and widens the other. What it points to instead is that these populations are far from equilibrium.

Approach

I develop analytical frameworks for the effects of linked and unlinked selection on neutral diversity, using moment equations and allele frequency spectra, and connect them to empirical measurements from monitored wild populations. Doing this credibly means accounting for recombination landscapes and functional target sizes from genomic data rather than assuming uniform parameters, since both strongly modulate how much diversity linked selection can remove.

A related strand extends this to stabilizing selection, where the existing toolkit is thin: most theory for linked selection is built around directional selection and sweeps, and a quantitative trait held near an optimum leaves a different footprint at linked sites.

Why it matters

Neutral diversity is the input to most demographic and conservation genetic inference — effective population size, historical bottlenecks, genetic health assessments. If selection at linked sites shapes that diversity more strongly than assumed, those inferences inherit the error. The same concern applies to demographic inference from the site frequency spectrum, where a high deleterious mutation rate can produce distortions that read as population expansion.

  • Mawass, Measured fitness variance is too weak to explain the missing diversity (in preparation)
  • Zhou, Mawass, Hayward & Berg, Linked neutral diversity under stabilizing selection: a two-locus moments model (in preparation)
  • Marcelle, Hernández, Mawass, Gutenkunst & Masel, High deleterious mutation rate distorts site frequency spectrum in mammals and leads to erroneous inference of demographic expansions (in preparation)