Robustness of the inference of the distribution of fitness effects to selective sweeps

fitdadi recovers the true distribution of fitness effects relatively well from selected-site SFS sampled immediately after fixation, but performance declines at lower recombination rates and with the unfolded SFS.

• Why it matters: DFE inference from the site frequency spectrum can be distorted by selective sweeps that hitchhike linked alleles, potentially violating assumptions about independence and shared demographic effects.
• What they did: The authors simulated genomic segments with neutral and deleterious mutations drawn from a known DFE under varying sweep conditions, then tested whether fitdadi with a flexible demographic model recovers that DFE.
• The result: With recurrent selective sweeps at rates consistent with human polymorphism and divergence data, the demographic model captures sweep distortions and yields an unbiased DFE, implying bias is unlikely but folded SFS and excluding low-recombination regions may be important.