Chromatin features define adaptive genomic regions in a fungal plant pathogen

Abstract

Understanding the complex information stored in a genome remains challenging since multiple connected regulatory mechanisms act at various scales to determine function. Increased comprehension of genome function at scales beyond contiguous nucleotides will help understand genetic diseases, the emergence of pathogenesis, and more broadly the genomics of adaptation. Here we report the analysis of DNA methylation, histone modification, and DNA accessibility in the plant pathogenic vascular wilt fungus Verticillium dahliae. Functional analysis details that DNA methylation is restricted to repetitive elements, such as transposable element DNA, but interestingly only some repetitive DNA is methylated. This incomplete DNA methylation is associated with repetitive DNA residing in specific compartments of the genome that were previously defined as Lineage-Specific (LS) regions. These regions are hypervariable between V. dahliae isolates and contain genes that support host colonization and adaptive traits. LS regions are associated with H3 Lys-27 methylated histones (H3K27me3), and repetitive DNA within LS regions are more transcriptionally active and have increased DNA accessibility, representing a hybrid chromatin state when compared to repetitive regions within the core genome. We used machine learning algorithms trained on epigenetic and DNA accessibility data to predict LS regions with high recall, identifying approximately twice as much LS DNA in the V. dahliae genome as previously recognized. Collectively, these results characterize LS regions in an intermediate chromatin state and provide evidence that links chromatin features and genome architecture to adaptive regions within the genome.