Yeast strain development has been essential for improving efficiency, flavour diversity, and quality of beer fermentation. Such efforts often rely on laborious in vitro screening experiments. However, with the increasing availability of large-scale ‘omics’ data sets, it may be possible to replace or complement such experiments with in silico screening. Here, we briefly review the genetics associated with various desirable and undesirable traits in brewing yeast, and demonstrate how recent genomics, transcriptomics, and proteomics data sets derived from the 1,011 yeast genomes project can be exploited for identifying strains with potentially desirable phenotypes. The discussed phenotypes are related to fermentation performance, formation of desirable flavours, and mitigation of off-flavours. Finally, we perform wort fermentations with five strains from diverse backgrounds, with diverse predicted phenotypes, to validate the in silico predictions. Most predicted phenotypes correlated well with the measured phenotypes, including formation of desirable compounds like isoamyl acetate and ethyl octanoate, as well as formation of undesirable compounds like 4-vinyl guaiacol, diacetyl, and ethanethiol. Together, the results indicate that utilizing large ‘omics’ data sets can be a very useful tool for both strain selection and development for beer fermentation, and naturally other food and beverage fermentations as well. Compared to more traditional in vitro screening, this has several benefits, including lower costs, more rapid results and possibility to include more strains. We hope this can inspire and yield improved and more diverse brewing strains to the industry.