Coelho-2017a

PhD thesis

Proteomics and the study of protein-protein interactions (PPIs) have become a trending research topic in the past decade. Thanks to the high-throughput experimental methodologies, the amount of data generated and uploaded to heterogeneous data repositories is increasing exponentially, presenting several research opportunities. Since experimental protein interaction identification techniques are expensive, time consuming, and require expert analysis, computational methods will prove crucial to tackle this issue. Several efforts have already been made regarding this problem, resulting in models for predicting intra-species interactions. Nevertheless, these models cannot be directly applied to inter-species PPI prediction. Inter-species PPIs are responsible or the basis of colonization and infection by bacterial pathogens and thus, their identification is crucial to prevent and treat disease states. This thesis proposes new computational models to predict inter-species PPIs, the impact of such associations to the human host, and associations between drugs and bacterial targets. The innovation of this proposal comes from the use of machine learning techniques tuned to biological problems, combined with complex biological network analysis. The obtained results will offer the means to better understand the associations between biological mechanisms and disease states, and aid the development of new diagnostic and therapeutic tools.