Membrane Protein Topology Prediction:
 A Bayesian Approach

Winner: Young Bioinformatician of the Year 2003

Protein structure prediction is a cornerstone of bioinformatics. Computational prediction is especially necessary for membrane proteins as physical methods of structure determination exhibit limited success. I address three different aspects of prediction: a new predictor, a meta-predictor and a method for membrane protein identification, all based on Bayesian Belief Networks (BBNs). Bayesian network probabilistic models provide a flexible and powerful framework for statistical inference, and learn model parameters from data. The goal of inference is to find the distribution of a random variable in the network conditioned on values of other variables in the network. BBNs can be used to efficiently estimate optimal values of model parameters from data. Another major advantage of BBNs is the ability to combine machine learning with expert opinions.