Meiotic recombination is a fundamental genetic and evolutionary process, initiated by the
deliberate infliction of numerous double strand breaks (DSBs) on the genome. In most mammals,
these DSBs are specified by PRDM9, which binds DNA through a zinc finger (ZF) array and
makes two histone modifications that together serve to recruit the DSB machinery. In these
species, the ZF binding affinity is rapidly-evolving. Intriguingly, PRDM9 is not only found in
mammals but throughout vertebrates, and may be directing meiotic recombination there too.
Despite its broad phylogenetic distribution, the gene has been lost independently many times; in
these cases, the determinants of DSB location are less well understood but are associated with
promoter features. We propose four analyses that address these gaps in our understanding: Aim
1. Do non-mammalian species with an intact PRDM9 use it to direct recombination? We will test
this hypothesis in corn snakes, a vertebrate species that carries a complete and rapidly-evolving
PRDM9. We will infer a genetic map from linkage disequilibrium (LD) data as well as by End-
seq, a recently-developed approach to assay meiotic DSB frequencies in the genome. We will also
collect genomic data about salient histone marks, chromatin accessibility and expression levels.
These data will help us to establish if PRDM9 is used to direct recombination. The generality of
our findings will be evaluated by building and examining an LD-based map in a fish species with
an intact PRDM9, the Northern pike. Aim 2. What mechanisms direct the location of DSBs in
species lacking an intact PRDM9? Here, we will focus on two vertebrates: zebra finches, which
(like other birds) lack PRDM9 entirely, and swordtail fish, which lack the two N-terminal
domains. We will combine existing LD-based genetic maps with data that we will collect on DSB
frequencies, salient histone marks, chromatin accessibility, and expression levels. We will then
ask which genomic features influence local recombination rates and if they also play a role in
species with an intact PRDM9. Aim 3. What genes co-evolve with PRDM9? We will test 246
candidate genes for their co-evolution with PRDM9 across the vertebrate phylogeny. As a
byproduct, we will make available a pipeline to identify orthologs of interest. Aim 4. What drives
the evolution of PRDM9 binding? To answer this question, we developed a generative model,
from PRDM9 binding to population dynamics. We will extend our model, notably to characterize
conditions for the loss of PRDM9, and test key predictions with genomic and comparative data.
Thus, we will combine population genetic, phylogenetic and experimental approaches in four vertebrate
species to learn how DSBs are localized in the genome and how and why the mechanism differs among taxa.