Natural genomic variation in rice blast genomes is associated with specific heterochromatin modifications

Eukaryotic genomes are not randomly organized, and often display patterns related to transcriptional activity, gene density, DNA replication, and chromosome position. The principal drivers of genome organization and how it influences evolution remains unclear. In plant pathogenic fungi, there are different descriptions of genome organization and we lack a framework to understand the mechanisms and evolutionary consequences of these organizational patterns. To further understand fungal genome organization and evolution, we utilized sequence data from 86 isolates of the rice blast fungus, Magnaporthe oryzae, and independently compared them to four new reference genome assemblies to identify DNA sequence variation and synteny. We also completed histone modification profiles using genome- wide chromatin immunoprecipitation-sequencing for the four new reference strains. The results show that the genomes are broadly organized around characteristics of euchromatin and heterochromatin. We report that repressive histone marks, tri-methylation of H3 Lysine-27 and H3 Lysine-9, H3K27me3 and H3K9me3 respectively, are associated with higher SNP and INDEL frequencies and higher SNP density. Over half of the densely grouped SNPs occur outside of transposable elements (TE), indicating the association between heterochromatin and DNA variation is not restricted to TEs. To further understand the association between heterochromatin and DNA variation, we calculated genome-wide silent SNP frequencies and discovered that silent SNPs occur more frequently in genes marked by H3K27me3. Specifically analyzing a group of 863 effectors, we also find that effectors that are variably present in the population are associated with H3K27me3. For effectors that are stably present in the population, the rate of silent SNPs is again significantly higher for those in H3K27me3 domains. Analyzing larger DNA segments, regions defined as euchromatic tended to be syntenic between strains, while heterochromatic regions were more often non-syntenic. We conclude that in M. oryzae, regions of heterochromatin are significantly associated with multiple types of DNA variation naturally occurring in the population, even types that should not be subject to selection.