A Data Analysis Pipeline for Identifying Periodic Processes during Drosophila Development

In this paper, we propose a computational pipeline that can be used to unearth periodic processes and their regulators in any organism along with the networks governing such periodicity. Our approach is based on mining periodic subgraphs from temporal gene networks. Specifically, we collected 30 time varying gene networks inferred from temporal expression profiles of 588 Drosophila genes, computed periodic subgraphs in those networks, and analyzed them in a number of ways in order to discover their biological significance. Our results show that the largest connected component in the periodic subgraphs as well as hub genes and dense subgraphs in that component are highly enriched in periodically active gene-functions. We also devised a way to find the regulators of these periodic functions. We show the superiority of our approach as compared to a baseline method that computes aperiodic subgraphs by showing that similar analysis on aperiodic subgraphs fails to find any periodic or specific gene function. To the best of our knowledge, this work is the first of its kind in the field of network biology.