An improved mitochondrial reference genome for Arabidopsis thaliana Col-0

Abstract

Arabidopsis thaliana remains the foremost model system for plant genetics and genomics, and researchers rely on the accuracy of its genomic resources. The first completely sequenced angiosperm mitochondrial genome was obtained from A. thaliana C24 (Unseld et al., 1997), and more recent efforts have produced additional A. thaliana reference genomes, including one for Col-0, the most widely used ecotype (Davila et al., 2011). These studies were based on older DNA sequencing methods, making them subject to errors associated with lower levels of sequencing coverage or the extremely short read lengths produced by early-generation Illumina technologies. Indeed, although the more recently published A. thaliana mitochondrial reference genome sequences made substantial progress in improving upon earlier versions, they still have high error rates. By comparing publicly available Illumina sequence data to the A. thaliana Col-0 reference genome, we found that it contains a sequence error every 2.4 kb on average, including 57 SNPs, 96 indels (up to 901 bp in size), and a large repeat-mediated rearrangement. Most of these errors appear to have been carried over from the original A. thaliana mitochondrial genome sequence by reference-based assembly approaches, which has misled subsequent studies of plant mitochondrial mutation and molecular evolution by giving the false impression that the errors are naturally occurring variants present in multiple ecotypes. Building on the progress made by previous researchers, we provide a corrected reference sequence that we hope will serve as a useful community resource for future investigations in the field of plant mitochondrial genetics.