Draft Assembly of the F2 European Maize Genome Sequence and Its Comparison to the B73 Genome Sequence: A Characterization of Genotype-Specific Regions

Abstract

Maize is well known for its exceptional structural diversity, including copy number variants (CNVs) and presence/absence variants (PAVs), and there is growing evidence for the role of structural variation in maize adaptation. F2 is a European maize line resulting from a long-term independent evolution relative to the reference American line B73. It also presents strong heterosis when crossed to American lines related to B73 or PH207, which has been instrumental for the development of hybrid breeding in Northern Europe. De novo genome sequencing of the French F2 maize inbred line revealed 10,044 novel genomic regions larger than 1 kb, making up 88 MB of DNA, that are present in F2 but not in B73 (PAV). This set of maize PAV sequences allowed us to annotate PAV content and to identify 395 new genes. We showed that most of these genes display numerous features that suggest they are either rapidly evolving genes or lineage-specific genes. Using PAV genotyping on a collection of 25 temperate lines, we also analyzed and provided the first insights about PAV frequencies within maize genetic groups and linkage disequilibrium in PAVs and flanking regions. The pattern of linkage disequilibrium within PAVs strikingly differs from that of flanking regions and is in accordance with the intuition that PAVs may recombine less than other genomic regions. As it was shown by several other authors, most PAVs are ancient, while we show that some are found only in European Flint material, thus pinpointing structural features that may be at the origin of adaptive traits involved in the success of this material. We conclude by some words on future directions.