Wang, H.; Nielsen, J.; Zhou, Y.; Lu, H.

The widely used model organism, Saccharomyces cerevisiae is ubiquitously present in a variety of natural and human-associated habitats. Despite extensive studies of this organism, the metabolic mechanisms driving its adaptation to varying niches remain elusive. We here leveraged the genome sequences of 1,807 S. cerevisiae strains for creating a high-quality pan-genome, which facilitates the comprehensive characterization of the genetic diversity among species. Utilizing the pan-genome to infer strain-specific genome-scale metabolic models (ssGEMs), enables quantitative predictions of physiological phenotypes and examining the metabolic disparities among all the S. cerevisiae strains. We further performed integrative analyses of fluxomic and transcriptomics, as well as constrained 907 ssGEMs by transcriptome data, which revealed the significant transcriptional regulation in the metabolism of certain amino acids at the population level. Additionally, comparing all 907 refined ssGEMs showed that S. cerevisiae strains from various ecological niches had undergone reductive evolution at both the genomic and metabolic network levels, as opposed to wild isolates. Finally, multidimensional integrative analyses of the pan-genome, transcriptome, and metabolic fluxome revealed marked metabolic differences among S. cerevisiae strains originating from distinct oxygen-limited niches, such as human gut and cheese environments, and identified convergent metabolic evolution, such as downregulation of oxidative phosphorylation pathways. Collectively, our study develops large-scale strain-specific genome-scale models at the population level, presenting unprecedented opportunities for understanding yeast adaptive evolutionary mechanisms and yeast physiology research.