The aim of this ERC PoC proposal was to optimize and validate a state-of-the-art pipeline, called EnhancerDesign, to design small, strong and cell-type specific human enhancers for gene therapy. This pipeline consists of four main steps that were all addressed in this project: 1) cell type profiling using single-cell multi-omics, 2) topic modeling and traininggeneration of an enhancer model using deep learning, 3) design and creation of synthetic enhancers and 4) experimental validation of enhancer activity. Our work was performed in three different contexts: 1) state specific melanoma, 2) hepatocyte subpopulations and 3) brain cell types.
This POC grant allowed us to develop new tools for performing massively parallel reporter assays (MPRA) in more versatile conditions with the use of viral vectors or combined with cell sorting. We succesfully developed new computational methods for the generation of fully synthetic enhancers. Finally, we applied our full pipeline to a new target cell type, the hepatocyte, and decoded the regulatory grammar allowing the control of expression level and zonation. All together, these technological advances provide a strong foundation for our licensing/patenting opportunities.
