Systematic generation of mechanistic hypotheses linking signaling and metabolism in cancer with COSMOS

Mentor: Aurelien Dugourd

Overview

Cancer is a very heterogeneous disease. Depending on the tissue of origin, it can present very diverse molecular profiles. Understanding the context dependent crosstalk happening between signaling and metabolism in various cancer contexts can be very valuable to better understand this disease. The NCI60 is a data-set where 58 cancer cell lines were used to generate transcriptomic and untargeted metabolomic profiles. These cell lines cover various cancer origins, such as clear cell Renal carcinoma (786-0), ovary carcinoma (OVCAR). While extensive studies have been performed on the transcriptomic and metabolomic datasets of the NCI60, exploration of mechanistic connections between the two layers has never been done. 

For doing this project you will use the transcriptomic and metabolomic dataset of the 2-3 selected cell lines of your choice. The NCI60 dataset is made available to trainees as preprocessed normalised gene expression, metabolite abundance and transcription factor (TF) activities. These three types of data will be used as input for the cosmos R package. To do so, you will adapt a template script to choose a cell line, and explore different values for cosmos parameters to generate networks of mechanistic hypotheses connecting TFs with metabolites specifically deregulated in a given cancer cell line. Once you have obtained a network of mechanistic hypotheses for each of your selected cell lines, you will compare the identified TF/metabolite connections between cell lines, and search literature to find potential explanations of your cell line exclusiveness. The exclusive connections of different cell lines may serve as a base to propose potential metabolic vulnerabilities to exploit with specific treatments.

Software requirements:

  • R >= 4.1
  • cosmosR: devtools::install_github(“saezlab/cosmosR”)
  • carnival 1.3: remotes::install_github(“saezlab/CARNIVAL@b3a84c6ba9706547caca02644566d75ee621f568”)

Training materials:

Suggested reading: