We present the 207 Mb genome sequence of the outcrosser Arabidopsis lyrata, which diverged from the self-fertilizing species A. thaliana about 10 million years ago. It is generally assumed that the much smaller A. thaliana genome, which is only 125 Mb, constitutes the derived state for the family. Apparent genome reduction in this genus can be partially attributed to the loss of DNA from large-scale rearrangements, but the main cause lies in the hundreds of thousands of small deletions found throughout the genome. These occurred primarily in non-coding DNA and transposons, but protein-coding multi-gene families are smaller in A. thaliana as well. Analysis of deletions and insertions still segregating in A. thaliana indicates that the process of DNA loss is ongoing, suggesting pervasive selection for a smaller genome.

Summary The genome sequence of the plant model organism Arabidopsis thaliana was presented in December of the year 2000. Since then, the 125 Mb sequence has revealed many of its evolutionary secrets. Through comparative analyses with other plant genomes, we know that the genome of A. thaliana, or better that of its ancestors, has undergone at least three whole genome duplications during the last 120 or so million years. The first duplication seems to have occurred at the dawn of dicot evolution, while the later duplications probably occurred <70 million years ago (Ma). One of those younger genome‐wide duplications might be linked to the K‐T extinction. Following these duplication events, the ancestral A. thaliana genome was hugely rearranged and gene copies have been massively lost. During the last 10 million years of its evolution, almost half of its genome was lost due to hundreds of thousands of small deletions. Here, we reconstruct plant genome evolution from the early angiosperm ancestor to the current A. thaliana genome, covering about 150 million years of evolution characterized by gene and genome duplications, genome rearrangements and genome reduction.