A systems biology approach to solving the puzzle of unknown genomic gene-function association using grid-ready SVM committee machines

Genomic researchers face the common challenge of deriving the functions of genes and proteins from high-throughput data. Experimental validation of protein function is costly and time-consuming. With the increased effectiveness of computational intelligence approaches, researchers aim to target the problem with in silico prediction of protein interactions and functions. We propose a systems biology approach that consists of machine-learning and visualization intelligence and aims to predict protein-protein interactions and enhance protein function annotation. Our machine-learning intelligence, SVM committee machines, is compatible with grid computing and large-scale data analysis. In this paper, we not only elucidate the computational power of protein interactions prediction, but also aim to emphasize the interpretation of protein function annotation through protein interaction network analysis.