Understanding the functional consequences of genetic variation on gene regulation

Understanding the relationship between the non-coding sequence of the genome and the gene expression is one of fundamental goals of regulatory genomics. Perturbing certain locations in the non-coding DNA causes a disturbance to the precise spatial and temporal expression of the genes. Gene regulatory mechanisms determine the amount of change in the gene expression from variations in the sequence. Mathematical modeling of gene expression has been proven to be successful to establish a sequence-to-function relationship in a context aware manner and provide mechanistic explanation of the gene regulatory processes. In this thesis, we aspire to provide tools for understanding sequence-level encoding of gene regulation by applying thermodynamics-based models. More specifically, we provide a probabilistic framework to develop deeper insights about current knowledge of a gene’s regulatory mechanisms, objectively characterizing what a new experiment adds to such knowledge and quantifying how ‘informative’ that experiment is. In order to elucidate mechanisms of transcriptional regulation we use single nucleotide polymorphism data to further investigate different mechanistic hypotheses and provide knowledge of systems-level processes. We construct a probabilistic model to leverage our knowledge of transcriptional regulatory networks and identify variations that lead to a significant change. Through this work we not only advance the field of regulatory genomics, but potentially provide a path for identifying variations in the DNA that significantly effect phenotype and lead to a disease