Pan-Genomics in Brachypodium and Implications for Related Grasses

Most commonly, intra-species variation is characterized as a set of SNPs and small indels identified by comparing short read data to a single reference genome. In contrast, pan-genomics compares full genome assemblies and gene annotations from multiple individuals of a single species. This approach captures highly divergent and novel sequences not found in the reference genome. Pan-genomics has revolutionized the study of bacterial species, however, few studies have taken this approach with plants, presumably due to the large size and difficulty in accurately assembling plant genomes. To evaluate the feasibility and benefits of plant pan-genomics, we sequenced 54 inbred lines of the model grass, Brachypodium distachyon, to a median depth of 92x (10 lines were sequenced with additional 4-6kb mate-pair libraries for scaffolding). Analysis of whole genome de novo assembly and gene annotation reveals that it is possible to recover nearly full genomic sequences from each line, with up to 97% of the whole genome sequence included into syntenic psuedomolecules. Analysis of this data yields a high-confidence B. distachyon pan-genome that includes 13,408 core gene clusters found in all lines, 7,283 soft-core genes cluster absent from a few lines, and 17,195 shell gene clusters found in 3 to 52 lines. We find 30,691 gene clusters represented by the reference genome/reference control and a varying number of other genomes. In addition, we identify 7,135 gene clusters not represented in the reference line or controls but present in multiple divergent lines. We show that non-core genes are expressed at lower levels, have narrower and more variable expression across accessions, are evolving faster, have reduced orthology to related grasses and are less likely to have a homeologous gene retained from the ancient genome duplication in the grass lineage. We evaluate the relationship between the number of sequenced lines and their phylogenetic position in relation to the addition of both genic and non-coding sequence. We describe the physical chromosome position of non-core and non-reference genes and its relation to transposable elements. This analysis suggests possible mechanisms by which dispensable genes are eliminated and also barriers to their removal. All lines sequenced in this study are available to researchers.