Extreme purifying selection against point mutations in the human genome

Area of Science:

• Human genomics
• Population genetics
• Evolutionary biology

Background:

• Large-scale genome sequencing advances allow measuring purifying selection in protein-coding genes.
• Selection in noncoding regions remains less understood due to methodological challenges.

Purpose of the Study:

• Introduce ExtRaINSIGHT, a novel method for quantifying ultraselection in both coding and noncoding human genomic regions.
• Assess the genome-wide distribution of strong purifying selection and its implications for deleterious mutation rates.

Main Methods:

• Developed ExtRaINSIGHT to estimate ultraselection by analyzing the depletion of rare single-nucleotide variants.
• Controlled for mutation rate variations to accurately measure selection.
• Applied the method to 71,702 whole genome sequences from the gnomAD v3 dataset.

Main Results:

• Identified abundant ultraselection in evolutionarily ancient microRNAs (miRNAs) and neuronal protein-coding genes.
• Found significant ultraselection at splice sites, but less in other noncoding RNAs and transcription factor binding sites.
• Estimated that approximately 0.4-0.7% of the human genome is ultraselected.

Conclusions:

• Ultraselection is prevalent in functionally critical genomic regions, including ancient miRNAs, neuronal genes, and splice sites.
• The findings suggest a substantial load of strongly deleterious mutations per generation (0.26-0.51).
• This study provides a comprehensive view of genome-wide fitness effects by integrating deep sequencing data with population genetics principles.