Predicting HIV-1, Host Protein-Protein Interactions
  • FENS
  • Predicting HIV-1, Host Protein-Protein Interactions

You are here

2012, Sabanci University

Title: Predicting HIV-1, Host Protein-Protein Interactions 


New infectious viruses appear regularly, established ones fail to be eradicated, posing significant challenges to public health. The lack of understanding of the intimate relationship between viruses and their hosts makes it difficult to develop effective therapies to treat viruses. Interactions between viruses and host proteins are key in this relationship, where viruses exploit the host proteins in order to hijack the cellular machinery. Considerable progress has been made with HIV-1, the causative agent of AIDS, in identifying thousands of physical interactions and functional associations between the virus and the human host proteins through experimental research efforts. However, the complete and accurate repertoire of the physical interactome is still far from complete. In order to better define the virus-host interactome, our work compliments experimental efforts by bridging different levels of biological information in a machine learning framework. A wide array of genomic and proteomic data serving as direct and indirect evidence for virus and host PPIs was integrated in a supervised learning framework. The predicted virus-host interactions were used to guide experimental studies by reducing the hypothesis space of all possible interactions to a tractable set.


If time permits I would also discuss our semi-supervised learning model for predicting HIV-1 host interactions and our efforts to obtain high-quality labels of protein-protein interactions.

 Short Bio:Oznur Tastan is a post-doctoral research fellow at Microsoft Research New England (Cambridge, MA, USA) since August 2010. She completed her PhD in Computer Science at Carnegie Mellon University under the supervision of Profs Jaime Carbonell and Judith Klein-Seetharaman. Oznur received her BSc in Biological Sciences at Sabanci University in 2004. Her research interests include machine learning and computational biology with a recent focus on cancer systems biology.