Investigating uncharacterised genes in Saccharomyces cerevisiae using robot scientists

Despite extensive research on Saccharomyces cerevisiae functional genomics, approximately 880 out of ~6,000 open reading frames (ORFs) remain uncharacterised. In this study we propose a method for characterising genes with limited prior functional knowledge using an automated laboratory platform, in conjunction with several hypothesis instantiation methods. We demonstrate this method by investigating YGR067C, an uncharacterised ORF hypothesised to regulate respiration during the diauxic shift. Predictions of the first-order effects of deletion were obtained by curating a list of pathways relevant to the hypothesis. Higher-order effects were predicted using simulation models based on the GEM Yeast9. The predictions were tested using empirical data from biological experiments performed in the Robot Scientist Eve, which generated OD560, transcriptomics, and metabolomics data. We observed that YGR067C deletion led to downregulation of transcripts in some ethanol consuming respiratory pathways during the glucose phase During the ethanol phase we observed that NAD+, NADP+ and NADH accumulated, and several amino acid biosynthesis pathways were enriched for the ygr067c∆ strain, suggesting longer term consequences of YGR067C mediated regulation. Based on these observations we propose that the role of YGR067C during the diauxic shift is to regulate genes related to ethanol consumption and respiration in the glucose phase.