Machine Learning methods to identify Spatial Gene Networks and understand tumor-microenvironment interactions of pituitary adenomas – SpaceTranscriptomiX

Single-cell sequencing and spatial transcriptomics have the potential to revolutionize our understanding of complex biological systems, but methodological challenges remain before we could exploit their full potential. Standardized methods are indeed required to build spatial signaling networks and identify the ecological cell-cell interactions that take in consideration spatial location and the functional mechanisms that drive communications. As a model of ecological system, the Tumor MicroEnvironment (TME) represents an intricate physical/signaling network that orchestrate exchanges between tumor-cells and cells of their TME (immune cells, endothelial cells, fibroblasts, …). SpaceTranscriptomiX focuses on the TME of pituitary adenomas (PAs) that are endocrine benign neoplasms (15% of intracranial tumors) which invasive progression is frequent and causes important co-morbidities due to their difficult surgical resection. This project aims at building a comprehensive map of the cellular and molecular networks that exist between pituitary tumor cells and the cellular components of their TME to decrypt the mechanisms of PAs invasive behavior. These objectives will be addressed through a tight collaboration between machine learning and PAs experts and the development of new methodological approaches. The project will: i) integrate expression and imaging data to decipher the spatial biological heterogeneity of PAs, ii) develop machine learning approaches to characterize spatial gene expression signatures and cell-cell communication networks, iii) validate the relevance of the developed pipelines to model the ecological interactions between Tumor and TME cells that underlie the invasive behavior of PAs. SpaceTranscriptomiX should deliver new machine learning methods to overcome the challenges to model the networking interactions from Spatial transcriptomic that will serve the understanding of PA invasive behavior and benefit to scientists using single-cell technologies.