Genetic variants have long been established as important contributors to human disease. In the past decade, a massive amount of human genetic variants, both coding and non-coding, have been discovered owing to the rapid development of next-generation sequencing technologies. On the other hand, the development of drugs for treating human diseases have also become increasingly dependent on computational methods. In this dissertation, I present several methods aimed at disentangling the functional impact of human genetic variants as well as expanding our knowledge about the effects of drugs on human bodies through network-based approaches and machine learning. In Chapter 2, I describe the construction of a three-dimensional regulatory network that generates mechanistic insight into the functional impact of both coding and non-coding disease-associated mutations. In Chapter 3, I describe an ensemble machine learning algorithm for predicting protein-protein interaction interfaces and how it spurs functional genomic studies. In Chapters 4 and 5, I propose machine learning methods for predicting drug- drug interactions and drug targets through biological networks. In Chapter 6, I describe a statistical method and a web tool for prioritizing variants of uncertain significance (VUS) leveraging spatial genomic patterns.