142 telomere-to-telomere assemblies reveal the genome structural landscape inSaccharomyces cerevisiae

Abstract

As population genomics is transitioning from single reference genomes to pangenomes, major improvements in terms of genome contiguity, phylogenetic sampling, haplotype phasing and structural variant (SV) calling are required. Here, we generated the Saccharomyces cerevisiae Reference Assembly Panel (ScRAP) comprising 142 reference-quality genomes from strains of various geographic and ecological origins that faithfully represent the genomic diversity and complexity of the species. The ca. 4,800 independent SVs we identified impact the expression of genes near the breakpoints and contribute to gene repertoire evolution through disruptions, duplications, fusions and horizontal transfers. We discovered frequent cases of complex aneuploidies, preferentially involving large chromosomes that underwent large SVs. We also characterized the evolutionary dynamics of complex genomic regions that classically remain unassembled in short read-based projects, including the 5 Ty families and the 32 individual telomeres. Overall, the ScRAP represents a crucial step towards establishing a high-quality, unified and complete S. cerevisiae pangenome.

Experimental approaches to study evolutionary cell biology using yeasts

“…While traditionally considered suboptimal for the imaging of intracellular features, super‐resolution techniques were recently developed (Chen et al, 2021; Hinterndorfer et al, 2022; Korovesi et al, 2022), and used (Dey et al, 2020) in yeast to study cellular processes at higher resolution, and new tools have been provided for the imaging of several subcellular structures (Akhuli et al, 2022; Zhu et al, 2019). A large number of telomere‐to‐telomere assemblies recently revealed the landscape of structural variants of S. cerevisiae natural isolates (O'Donnell et al, 2022). Finally, synthetic biology approaches such as The Synthetic Yeast Genome Project (S. M. Richardson et al, 2017) continue to provide novel tools to generate synthetic cellular conditions.…”

VRPG: an interactive visualization framework for reference-projected pangenome graph

“…With the SGDref as the reference genome, we incrementally added those 162 ScRAP assemblies into the graph according to their phylogenetic distances to SGDref. The phylogenetic distances employed here were extracted from the phylogenetic tree of these 163 input genomes built upon their concatenated 1-to-1 orthologous gene matrix (See [16] for details). Regarding the haplotype tag, for the yeast reference pangenome graph, we used “HP0” to denote haplotypes of haploid or homozygous diploid strains, while using “collapsed”, “HP1”, and “HP2” to denote collapsed, or the two phased haplotypes of heterozygous diploid strains.…”

“…To demonstrate the application of VRPG in real world examples, we set up a demonstration website (https://www.evomicslab.org/app/vrpg/) to visualize reference pangenome graphs derived from 143 yeast and 90 human genome assemblies respectively. The human reference pangenome graph is a public dataset, whereas the yeast reference pangenome graph is built based on the Saccharomyces cerevisiae reference assembly panel (ScRAP) that we recently constructed (O’Donnell et al ., 2022). Both graphs were constructed by Minigraph (Li et al ., 2020) via a standard reference pangenome graph building protocol (Supplementary Note).…”

The genomic landscape of transposable elements in yeast hybrids is shaped by structural variation and genotype-specific modulation of transposition rate

“…While producing collapsed assemblies for haploid or predominantly homozygous genomes is relatively straightforward, complex heterozygous genomes are much more difficult to characterize at the haplotype level. Recent studies demonstrated the feasibility of the approach (Heasley and Argueso, 2022; O’Donnell et al, 2023), but the task of resolving hybrid genomes across a wide range of parental divergence under a unified framework remains challenging. Here, we employ a phasing and assembly approach using long reads to resolve hybrid MA lines genomes.…”