UC San Diego

Transcriptional gene regulation is a primary mechanism that Escherichia coli uses to best adapt to its current environment. RNA-sequencing data in particular provides us with the ability to more clearly examine the different states of the transcriptome and thus study transcriptional regulation. The cost to generate RNA-sequencing datasets has dramatically decreased over the last two decades which, in conjunction with FAIR data principles, has subsequently led to a large increase in the amount of publicly available RNA-sequencing data. In addition, centralized public databases of both these large datasets and biological knowledge have become increasingly large and accessible. Both developing and deploying new analytical techniques are necessary in order to best derive actionable insights from this large increase in scale for biological research. In this thesis, we first apply these principles to studying RNAP mutations and their ability to shift a transcriptome to favor growth over stress functions. This tradeoff between fear and greed can be seen in nearly all RNAP mutations and also across numerous bacterial species. Next we investigate the plasticity of transcriptional regulation by removing selected transcription factors and evolving strains, finding some knockouts recover growth without significant adaptation while others require convergent mutations to regulatory elements which restore the expression of highly growth-important genes. Finally, we build a model for the transcriptional regulatory network using iModulons as a measure of regulator activity. This model is able to accurately predict metabolite concentrations as well as infer biological constants about transcription factor binding. Altogether, these projects help advance our understanding of the mechanisms underlying transcriptional regulation through the utilization of data science.