Analysis of 700,000 exomes refines mutation rate estimates and variant effect inference

A preprint posted to bioRxiv on 10 June 2026 describes a large-scale analysis of coding variation using the gnomAD v4 dataset, which contains exome sequences from approximately 730,947 individuals — the largest publicly available exome resource to date. The authors focus on two related problems: estimating locus-specific mutation rates more accurately by exploiting the high density of independently recurrent rare variants that gnomAD v4 makes visible, and improving inference of variant effects from this enriched dataset.

The study is motivated in part by the growing clinical deployment of genomic sequencing, including its emergence as a component of newborn screening programmes, which creates demand for comprehensive and accurate annotation of rare Mendelian disease genes and their pathogenic variants. Accurate mutation rate models are foundational to variant classification frameworks and to distinguishing variants of uncertain significance (VUS) from likely pathogenic findings.

By exploiting independent recurrence of rare coding variants — events that become tractable at gnomAD v4 scale — the authors argue that locus-specific mutation rates can be inferred at higher resolution than previously possible. The preprint has not yet completed peer review and analytical details require independent evaluation.