Closing and Finishing the Mitochondrial and Chloroplast Genomes of Sorghum Male-Sterile a1 Cytoplasm Using Long Span NGS Read Technology

To gain further insight into cytoplasmic-male sterility (CMS) and the action of fertility restoration genes (Rf) in sorghum, we sequenced and assembled the A₁ CMS mitochondrial genome from a subcellular fraction enriched in mitochondria and chloroplasts. De novo sequence assembly of these large, repeat rich genomes is complicated by significant lateral gene transfer between the organelles. Using a new long span library construction strategy we were able to assemble both genomes via an 8Kb mate pair (MP) and a 35-50Kb MP library using 2x250 read chemistry. Data from the MP libraries were processed to filter out chimeric reads using a novel encrypted barcode strategy. Raw deduplicated fragment and MP reads were filtered by mapping to the reference genomes of normal-cytoplasm mitochondria (NC_008360.1; 468,628 bp) and chloroplasts (EF115542.1; 140,754 bp). Approximately 80% of fragment reads mapped to the mitochondrial genome, 16% to the chloroplast genome, and 4% were found to match nuclear sequences. De novo assembly was performed using non-chloroplast fragment reads to assemble the mitochondrial genome, and non-mitochondrial fragment reads to assemble the chloroplast genome. Resulting contigs were then mapped with similarly filtered 8Kb MP data and assembled into scaffolds, and the resulting scaffolds were closed by mapping of 35-50Kb MP data. The mitochondrial genome assembled into a circular molecule of 454Kb, and exhibits multiple rearrangements, indels and duplications relative to the reference strain genome. This new long span NGS read technology permits assembly of complex genomes, and enabled detection of potential multiple isoforms of the mitochondrial genome.