A new evolutive clustering algorithm (eSIMBA) for active module identification in p-value attributed temporal biological networks

Biological processes are inherently dynamic: gene expression and molecular interactions fluctuate as cellular mechanisms evolve over time. Active Module Identification Methods (AMIMs) capture these functional behaviours by identifying sets of molecular entities exhibiting coordinated activity. Yet, understanding biological systems requires more than a static picture and calls for the temporal contextualization of such modular dynamics. To address this need, the proposed algorithm combines static and sliding-time-window community detection using a greedy matching procedure that maintains continuity across time. This method detects active modules at each time point and consistently tracks their evolution. Extensive experiments performed on both synthetic and real-world biological networks confirm its effectiveness. By explicitly embracing temporal variability, the proposed approach reveals how active modules emerge, split, merge, or disappear, providing new insights into the dynamic organization of node-attributed biological networks.