Paper detail

On prediction of regulatory genes by analysis of C.elegans functional networks

Connectivity networks have recently become widely used in biology due to increasing amounts of information on the physical and functional links between individual proteins. This connectivity data provides valuable material for expanding our knowledge far beyond the experimentally validated via mathematical analysis and theoretical predictions of new functional interactions. In this paper we demonstrate an application of several algorithms developed for the ranking of potential gene-expression regulators within the context of an associated network. We analyze how different types of connectivity between genes and proteins affect the topology of the integral C.elegans functional network and thereby validate algorithmic performance. We demonstrate the possible definition of co-expression gene clusters within a network context from their specific motif distribution signatures. We also show that the method based on the shortest path function (SPF) applied to gene interactions sub-network of the co-expression gene cluster, efficiently predicts novel regulatory transcription factors (TFs). Simultaneous application of other methods, including only interactions with neighborhood genes, allows rapid ranking of potential regulators that could be functionally linked with the group of co-expressed genes. Predicting functions of regulators for a cluster of ribosomal/mRNA metabolic genes we highlight a role of mRNA translation and decay in a longevity of organisms.