In this dissertation, we tackle problems in gene regulation and distance metric learning. In the first part of this thesis, we present three novel approaches for modeling transcriptional and post-transcriptional gene regulatory mechanisms. First, we propose a causal reasoning model for inferring upstream regulators of gene expression, including transcriptional regulators. Second, we propose a model for predicting small RNAs (sRNAs) in bacterial species that act as post-transcriptional regulators of the global regulator CsrA. Third, we propose a generalization of genome-wide association study (GWAS) over regulatory networks to identify functional pathways that are associated with a complex trait. Finally, in the second part of this thesis, we present a reformulation of the distance metric learning problem. All of our methods achieve good performance, are computationally efficient and are implemented in open-source R packages which can be installed from public repositories.