Mots-Clés

transcriptomique, métabolomique, networks

Description

Leveraging the metabolomic and transcriptomic public resources to study the link between gene expression and metabolic programs

There is correlation between transcriptomic and metabolomic profiles, but this relationship is complex and depends on many factors to which transcriptomics is blind. For instance, enzymes can be regulated by posttranslational modifications and protein abundance is not always predicted accurately by transcript abundance.

However, a number of software tools have been proposed to infer metabolic pathway activity on the basis of bulk or single-cell transcriptomic profiles (COMPASS, scFEA, METAFLux…).  These tools are based on graphs that integrate all the available information about metabolic reactions in an organism and the genes that encode each enzyme. To our knowledge, no benchmark has been performed yet on these newly developed methods despite the practical importance of this task. We are especially interested in testing such software in the context of solid tumors, including their microenvironment, be it at single-cell resolution or in bulk.

Our team has a recognized expertise in the development of tools based on networks (1-2). We propose an internship to realize a joint analysis of metabolite usage and gene expression at a broad level using more than 150 cancer cell lines and a vast repository of public tumors from the Cancer Atlas of Metabolic Profiles (3) (764 tumor samples and 224 adjacent normal samples, across 11 different cancer types, covering 15 datasets for both metabolomic and transcriptomic). The metabolomic dataset will be used as ground truth to benchmark methods previously mentioned based on transcriptomics.

This work will serve as a basis to decipher the link between metabolism and transcriptomic dataset, and to compare publicly available tools with internal developments of the team. The student will notably investigate the case of A375 melanoma cells resistant to BRAF inhibitor vemurafenib and treated with fatty acid oxidation inhibitor ranolazine where both transcriptomic and metabolomic data are available (4). Successful work could lead to a benchmark paper.

References
1- Cabello-Aguilar S & al.  Alame M, Kon-Sun-Tack F, Fau C, Lacroix M, Colinge J. SingleCellSignalR: inference of intercellular networks from single-cell transcriptomics. Nucleic Acids Res. (2020)
2 - Villemin &  al. Inferring ligand-receptor cellular networks from bulk and spatial transcriptomic datasets with BulkSignalR, Nucleic Acids Res. (2023)
3 - Benedetti & al; al A multimodal atlas of tumor metabolism reveals the architecture of gene–metabolite covariation. Nature Metabolism (2023)
4 - Redondo-Muñoz & al.  Using Metabolic rewiring induced by ranolazine improves melanoma responses to targeted therapy and immunotherapy. Nature Com. (2023)