Long-read sequencing reveals pre-meiotic gene conversion in sperm

Abstract

Meiotic recombination is a fundamental process that generates genetic diversity by creating new combinations of existing alleles1. Whereas crossovers in humans are well characterized2, the more frequent non-crossovers that lead to gene conversion remain challenging to study. Here we show that single high-fidelity long sequencing reads from sperm can capture both crossovers and non-crossovers, which enables effectively arbitrary sample sizes for analysis from a single male. We analysed 2,382 candidate non-crossovers in 15 sperm samples from 13 donors, and identified a consistent component with properties distinct from PRDM9-induced recombination. This phenomenon was not associated with meiotic double-strand break sites identified by DMC1 binding, the crossover recombination map or GC-biased gene conversion, but was associated with genomic fragile sites. This component is also seen in paternal non-crossover gene conversions in pedigree data3. Applying the same analysis to 12 blood samples4, we observed non-crossover gene conversions with similar properties, but very few crossover events. Further, we demonstrate variation between donors for the different types of recombination, even when they share the same PRDM9 genotype. We suggest that a substantial fraction of the non-crossover gene conversion events seen in sperm arise prior to meiosis. Single-molecule long-read sequencing of human sperm reveals variation in recombination across donors and that a substantial fraction of non-crossover gene conversions arises before meiosis, distinct from PRDM9-driven recombination and associated with fragile sites.