Chromosome-level assemblies of multiple Arabidopsis genomes reveal hotspots of rearrangements with altered evolutionary dynamics

Abstract

Despite hundreds of sequenced Arabidopsis genomes, very little is known about the degree of genomic collinearity within single species, due to the low number of chromosome-level assemblies. Here, we report chromosome-level reference-quality assemblies of seven Arabidopsis thaliana accessions selected across its global range. Each genome reveals between 13-17 Mb rearranged, and 5-6 Mb non-reference sequences introducing copy-number changes in~5000 genes, including~1900 non-reference genes. Quantifying the collinearity between the genomes reveals~350 euchromatic regions, where accession-specific tandem duplications destroy the collinearity between the genomes. These hotspots of rearrangements are characterized by reduced meiotic recombination in hybrids and genes implicated in biotic stress response. This suggests that hotspots of rearrangements undergo altered evolutionary dynamics, as compared to the rest of the genome, which are mostly based on the accumulation of new mutations and not on the recombination of existing variation, and thereby enable a quick response to the biotic stress.

GALA: gap-free chromosome-scale assembly with long reads

“…5). Of note is that this failure also occurred in most of other commonly used computational tools too [19,[22][23][24].…”

The evolutionary dynamics of genetic incompatibilities introduced by duplicated genes inArabidopsis thaliana

“…The eight AMPRIL founder accessions (An-1, C24, Col-0, Cvi-0, Eri-1, Kyo, Ler and Sha) were selected across the entire geographic distribution of A. thaliana including the Northern hemisphere and the Cape Verde Islands. Recently, we generated chromosome-level genome assemblies of all seven, non-reference founder genomes 22 . The first release of the AMPRIL population (AMPRIL I) contained six subpopulations (referred to as ABBA, ACCA, ADDA, BCCB, BDDB, CDDC) derived from reciprocal diallel crosses between four hybrids (A: Col-0 x Kyo, B: Cvi-0 x Sha, C: Eri-1 x An-1, D: Ler x C24) and the subsequent selfing to the F4 generation by single-seed descent 21 .…”

“…We used previously released whole-genome short read data 22 to generate markers for genotyping the AMPRILs. We mapped the reads to the reference sequence and called SNPs using SHORE (version 0.…”

“…For each duplicated gene pair, we required two copies in the reference sequence and at least one copy in one of the other genomes (DupGene2), or one copy in reference sequence and at least one copy on a different chromosome in at least one of the other parental genomes (DupGene1). We assessed the genotypes of each of the gene copies in each AMPRIL using the genotypes predicted in the middle of the respective reference gene, or in the case a gene was not present in the reference sequence -using the midpoint between the two closest flanking syntenic regions of the non-reference gene copy (based on synteny calculations from a previous study 22 ).…”

Sequencing and analysis ofArabidopsis thalianaNOR2 reveal its distinct organization and tissue-specific expression of rRNA ribosomal variants

“…Although the ribosomal RNA transcripts account for approximately 50% of all transcribed RNA in a cell, only a fraction of the tandemly repeated rDNA genes is transcribed at a given time 6,7 . The transcriptional regulation of individual rRNA genes and their organization within the NOR domains remains unknown, for any organism, due to their repetitive nature that hinders classical sequencing approaches 7,10 . In fact, it has never been possible to assemble the complete NOR by directly sequencing the genome of Arabidopsis or that of any organism.…”