Inyoung Kim

Virginia Tech

Sparsity Priors for Protein-Protein Interaction Predictions

Protein-protein interactions play important roles in most fundamental cellular processes including cell cycle, metabolism, and cell proliferation. Therefore, the development of effective statistical approaches to predicting protein interactions based on recently available large scale experimental data is a very important problem. However, due to the number of protein-protein interaction to be observed is very small, the number of parameters to be estimated is very large. Therefore the data is very sparse due to a few number of protein-protein interaction to be observed. In this paper, we incorporate a point-mass mixture prior in the analysis through a Bayesian method. The prediction results between with and without this prior are compared using the large-scale protein-protein interaction data obtained from high throughput yeast two-hybrid experiments. The result demonstrates the advantages of the Bayesian approach with a sparsity prior based on point-mass mixture prior.