Structural and regulatory variation in the grapevine genome

Abstract

The grapevine genome harbors considerable levels of structural variation that represent an important source of functional variation and could result in high levels of cis-regulatory variation. For structural variation we focused on the elucidation of the the elusive structure of centromeres in grapevine species. Centromere specific haplotypes belonging to 3 cultivated grapevine accessions, one wild accession of the same species and one interspecific hybrid have been reconstructed using PacBio HiFi and ONT long-reads. ChIP-seq data of centromere specific histone variant CENH3 has also allowed us to identify the functional domains of the centromeres and study their evolution throughout the domestication process. Considerable variation in repeats arrays dimensions and repeat composition in centromeric haplotypes was present both within the same individual and between accessions. The variation mostly involved non-functional satellite repeats, namely the very abundant 107 bp repeat, while the functional part composed of small arrays of a 59-mer monomer and a solo-LTR Gypsy element was conserved overall when comparing the same chromosome in different accessions. The comparison of identical by descent haplotypes has also uncovered recent events of TE insertions and generation of solo-LTR, giving hints on the underlying mechanisms of diversification and homogenization of these regions.

For cis-regulatory variation, we developed a new theoretical and experimental approach to identify the cis-regulatory alleles present in a population of individuals and to estimate their cis-regulatory value that we named Estimated Genetic Contribution (EGC) of alleles. We have applied our approach to analyse transcript abundances in leaves and berries of >90 whole genome sequenced grapevine genotypes. We observe that the cis-regulatory variation of individual genes is mostly due to the presence of a multiallelic system with many different levels of expression related to the presence of multiple haplotypes in the population and that the EGC of individual haplotypes is highly variable between organs both in terms of the magnitude of relative differences among haplotypes (dynamic range of expression variation) as well as in terms of relative expression values of different haplotypes (ranking of the different haplotypes in terms of EGC). We also observe different levels of cis-regulatory variation between leaves and berries.