Transmission distortion affecting human noncrossover but not crossover recombination: A hidden source of meiotic drive

Author Summary Meiosis is an essential feature of sexual reproduction that maintains chromosome number over generations. This specialised form of cell division creates gametes containing a single copy of each chromosome so that each parent contributes half their genetic information to an offspring. Accurate partitioning first requires intimate association of the two parental copies of each chromosome and concomitant exchange between them. These exchanges consist of both large-scale reciprocal crossovers, essential for correct chromosome segregation, and very localised gene conversion events, or noncrossovers, thought to be involved in correct chromosome pairing. Ordinarily, the reshuffling of genetic variants between generations by recombination will not alter their population frequency only their haplotypic context, with a parent passing on a given variant to 50% of its gametes according to Mendel’s law of inheritance. However, by screening for both types of recombinant amongst the sperm DNA of healthy men, we have identified a novel form of biased transmission that is restricted to noncrossovers and favours eventual fixation of one variant over another in the population. This previously undetected source of meiotic drive will not alter recombination propensity but is likely to be a common and potent force acting on the human genome. Meiotic recombination ensures the correct segregation of homologous chromosomes during gamete formation and contributes to DNA diversity through both large-scale reciprocal crossovers and very localised gene conversion events, also known as noncrossovers. Considerable progress has been made in understanding factors such as PRDM9 and SNP variants that influence the initiation of recombination at human hotspots but very little is known about factors acting downstream. To address this, we simultaneously analysed both types of recombinant molecule in sperm DNA at six highly active hotspots, and looked for disparity in the transmission of allelic variants indicative of any cis-acting influences. At two of the hotspots we identified a novel form of biased transmission that was exclusive to the noncrossover class of recombinant, and which presumably arises through differences between crossovers and noncrossovers in heteroduplex formation and biased mismatch repair. This form of biased gene conversion is not predicted to influence hotspot activity as previously noted for SNPs that affect recombination initiation, but does constitute a powerful and previously undetected source of recombination-driven meiotic drive that by extrapolation may affect thousands of recombination hotspots throughout the human genome. Intriguingly, at both of the hotspots described here, this drive favours strong (G/C) over weak (A/T) base pairs as might be predicted from the well-established correlations between high GC content and recombination activity in mammalian genomes.