Network Bioinformatics: from data analyses to predictive models
Systems biology studies biological processes considered as networks of interacting entities, namely as complex systems. Because these systems are multi-scaled and dynamic, this approach promises an integrated vision of cell, tissue, organ and population functioning.
Aiming at identifying all the parts constituting the biological complex systems and their interactions, systematic 'omics' programs have been launched at several levels of organization, from genome to interactomes.
Meanwhile, the switch from the gene-centric conception to the more holistic perception of the biological process as a whole, operated. Therefore, nowadays various 'omics' data obtained in different conditions are rapidly accumulating in public repositories and constitute an unprecedented source of knowledge related to expression, macromolecular and regulatory interactions, or genetic perturbations.
Analyzing these high-throughput data is the current bioinformatics challenge and this represents our main field of investigations. Indeed, extracting relevant biological information from such data implies (i) solving network representation and visualization issues; (ii) developing novel algorithmic solutions for network partitioning; (iii) mining and identifying statistically significant signals (motifs in sequences and graphs, annotation terms, etc.); (iv) introducing high-throughput data into dynamic models. Our abilities and activities cover these aspects which, noticeably, open formal mathematical, computational and statistical questionings/problems nourishing our close collaboration with the ‘Mathematical Methods for Genomics’ group at IML.
On the biological side, our bioinformatics group provides the diverse competencies necessary to investigate gene, protein, cell, tissues and organ functions with a network perspective. For this, we organize and analyse high-throughput data with in-house original and dedicated methods to tackle particular biological questions related to (i) the deciphering of gene regulatory networks and (ii) the prediction of protein function. Most of these studies are directly linked to experimental projects performed at TAGC, and are positioned either upstream (when predicting from prior knowledge) or downstream (when analyzing newly generated data) of the experimental part.