Cataloging Human <i>PRDM9</i> Allelic Variation Using Long-Read Sequencing Reveals <i>PRDM9</i> Population Specificity and Two Distinct Groupings of Related Alleles

The PRDM9 protein determines sites of meiotic recombination in humans by directing meiotic DNA double-strand breaks to specific loci. Targeting specificity is encoded by a long array of C₂H₂ zinc fingers that bind to DNA. This zinc finger array is hypervariable, and the resulting alleles each have a potentially different DNA binding preference. The assessment of <i>PRDM9</i> diversity is important for understanding the complexity of human population genetics, inheritance linkage patterns, and predisposition to genetic disease. Due to the repetitive nature of the <i>PRDM9</i> zinc finger array, the large-scale sequencing of human <i>PRDM9</i> is challenging. We, therefore, developed a long-read sequencing strategy to infer the diploid <i>PRDM9</i> zinc finger array genotype in a high-throughput manner. From an unbiased study of <i>PRDM9</i> allelic diversity in 720 individuals from seven human populations, we detected 69 <i>PRDM9</i> alleles. Several alleles differ in frequency among human populations, and 32 alleles had not been identified by previous studies, which were heavily biased to European populations. <i>PRDM9</i> alleles are distinguished by their DNA binding site preferences and fall into two major categories related to the most common <i>PRDM9-A</i> and <i>PRDM9-C</i> alleles. We also found that it is likely that inter-conversion between allele types is rare. By mapping meiotic double-strand breaks (DSBs) in the testis, we found that small variations in <i>PRDM9</i> can substantially alter the meiotic recombination landscape, demonstrating that minor <i>PRDM9</i> variants may play an under-appreciated role in shaping patterns of human recombination. In summary, our data greatly expands knowledge of <i>PRDM9</i> diversity in humans.