Cancer bioinformatics and systems biology : J. Colinge

Axis 1: Inter- and intracellular network inference

The analysis of large-scale molecular biology data is often highly complex, due to the number of molecules measured and the limited number of samples involved. Contextualizing data using molecular interaction networks is a classic approach to facilitating the extraction of relevant and reliable information. We combine the development of methods for inferring such networks, as well as for embedding datasets in existing or newly determined networks.

At present, our efforts are mainly focused on predictive aspects. We have published and are continuing to build a series of tools designed to reconstruct cellular ligand-receptor networks characterizing the ecosystem of cells making up a tissue (Cabello-Aguilar, Nucleic Acids Res, 2020 ; Villemin, Nucleic Acids Res, 2023). These methods apply to bulk, single-cell and spatial transcriptomics data. They are available as R libraries and involve data integration, graph-based algorithms, and statistical modeling.


We are also working on methods for reconstructing molecular networks from perturbation data. To this end, we have developed a new formulation of the Modular Response Analysis (MRA) approach by revisiting the analysis of MRA equations and applying multilinear regression techniques (least squares, lasso, STEP, etc.). This has defined a whole new family of methods that are much less sensitive to noise and can model networks of 10-1000 nodes. Here too, R code has been made available (Jimenez-Dominguez, Sci Rep, 2020 ; Borg, Bioinformatics, 2023).



We also have strong expertise in the application and creation of methods for projecting and integrating data over networks (graphs) such as random walks (Dwivedi, Malaria J, 2017 ; Müllner, Mol Syst Biol, 2015 ; Blomen, Science, 2015 ; Rix, PLoS ONE, 2013 ; Burkard, PLoS Comput Biol, 2010). Developments in this area are on hold, but we'll be coming back to it shortly.


Axis 2: Functional systems biology

The lab is involved in the analysis of numerous data sets in collaboration with our biologist and clinician partners. This work naturally relies on the application of our own methods, but also of any relevant tools developed by the community. The model study generally consists of a small or medium-sized cohort of a rare tumor, or one that answers a particular question (response to a therapy, patient sub-type, access to a new technology, etc.).

Typically, our analyses aim to provide an informative description of the tumor entity, as well as criteria for stratifying patients, understanding a mechanism of resistance or non-response, or suggesting new therapeutic opportunities. Recent examples include salivary gland ductal carcinoma (Alame, Theranostics, 2020), colorectal liver metastases (Giguelay, Theranostics, 2022) and pancreatic adenocarcinoma (Souche, Endoscopy, 2021).


We have several studies of this type underway (breast, colon, ovarian, pancreatic and various blood cell cancers) with various partners, including the Turtoi, Maraver, Djiane and Gongora labs at IRCM, the translational research unit at ICM (Evelyne Crapez), and numerous clinicians, including William Jacot, Didier Pourquier, Thibault Mazard at the ICM, Valérie Costes-Martineau, Eric Assenat at Montpellier University Hospital, Francine Garnache-Ottou, Florian Renosi at EFS Besançon, Eric Tartour at APHP, Olivier Adotévi at Besançon University Hospital, etc.


Axis 3: Modelling

The lab has extensive experience in the application of numerical analysis and computational statistical methods.

Proteomics is a key application area for us (Colinge, Proteomics, 2003; Breitwieser, J Proteome Res, 2011). In this domain, we are currently working on modeling the dynamics (turnover) of proteins in vivo (Lehmann, Anal Chem, 2019) with the teams of Sylvain Lehmann and Christophe Hirtz (CHU Montpellier, IRMB and INM). We are starting research on modeling variation in protein complexes in collaboration with Serge Roche (IGMM), an old familiar theme (Stukalov, J Proteome Res, 2012; Varjosalo, Nat Methods, 2013; Huber, Nat Methods, 2015; Mellacheruvu, Nat Methods, 2015) that we are revisiting with enthusiasm.


Circulating DNA analysis is a field of application for machine learning methods that enable cancer detection from a blood sample (Tanos, Adv Sci, 2020). We are collaborating with Alain Thierry's lab(IRCM), which is a leader in the field.


Finally, in parallel with our projects in spatial transcriptomics, we are applying spatial statistics techniques to tumor nanomechanics, i.e., the variation of tumor tissue stiffness (Crapez, Sci Rep, 2022). Project in collaboration with the ICM translational research unit (Evelyne Crapez), Thibault Mazard (ICM clinician) and Christine Bénistant (Milhiet CBS lab).




© Institut de Recherche en Cancérologie de Montpellier - 2011 - Tous droits réservés - Mentions légales - Connexion - Conception : ID Alizés