Regulatory Regions in Gene Regulation as an Information Bottleneck

Biological systems need to process information in order to perform specific functions. In the context of gene regulation, regulatory regions process transcription factor signals in order for cells to differentiate towards correct fates. In the past, we have shown that the information bottleneck (IB) framework provides a useful framework for understanding regulatory binding site regions. Here, I will discuss two recent modifications to models on the one hand and the IB framework applied to transcription factor sensing on the other with the goal of coming closer to predictions for these regions. First, using two complementary models for clustering transcription factors at binding site sensors, we can study information transfer during early fly embryo development with local transcription factor clustering. We find that weak cooperativity or clustering can allow for maximal information transfer, especially about the relevant variable, and that weak clustering also allows the binding site sensors to achieve optimality consistent with the IB bound. Second, we investigate how optimal activation changes when multiple binding site elements can process information, and find that activation profiles consistent with IB optimality resemble gene expression profiles in the early fly embryo.