Interpretable deep learning for inferring cellular transcriptional and metabolic changes of genetic perturbations in cancer (PERMET2CAN)
The main objective of the project is to develop a new framework based on interpretable Deep Learning (DL) to jointly infer transcriptional outcomes and metabolic changes associated with genetic perturbations. The design of the DL model will be guided by a priori biological information and will take into account two simultaneous objectives. Single-cell RNA sequencing (scRNA‑Seq) data from perturbation experiments will be used. The proposed framework will allow identifying key cellular responses, transcriptional and metabolic, resulting from genetic perturbations. Understanding transcriptional changes to genetic perturbations is central to identifying genetic interactions in cancer. Similarly, understanding the role of metabolism in cancer not only provides insights into the fundamental biolog...