Belief propagation estimation of protein and domain interactions using the sum-product algorithm

We present a novel framework to estimate protein-protein (PPI) and domain-domain (DDI) interactions based on a belief propagation estimation method that efficiently computes interaction probabilities. This methodology uses experimental interactions, domain architecture and GO annotations to create a factor graph representation of the joint probability distribution of protein and domain pairs. We use bound structures as a priori evidence of domain interactions. These structures come from experiments documented in iPfam. The probability distribution contained in the factor graph is then efficiently marginalized with a message passing algorithm called the Sum-Product Algorithm (SPA). This method is compared against two other approaches: Maximum Likelihood Estimation and Maximum Specificity Set Cover. SPA performs better for simulated scenarios and for inferring high quality PPI data of Saccharomyces cerevisiae. This framework can be used to predict potential protein and domain interactio...

• Type: Journal
• Year: 2010
• Where: TIT
• Authors: Faruck Morcos, Marcin Sikora, Mark S. Alber, Dale Kaiser, Jesús A. Izaguirre
• Keywords: Domain Interactions, Education, Factor Graph, Probability Distribution, TIT 2010