Demystifying the mystery genome for alternative immunotherapy targets: combining artificial intelligence-based discovery of immunomodulatory endothelial cell targets and in vivo target validation.

Society invests large budgets in basic research, but (too) many results run ashore in the Valley of Death and are not translated in drug development/clinical treatment. This is partly due to insufficient target discovery/validation and suboptimal target selection. Yet, the function of a third of the human coding genome remains “mysterious” without annotation, representing a goldmine for target discovery, knowledge gain and drug development. My aim is to address this problem by a multi-disciplinary approach, relying on synergy between artificial intelligence (AI) and biology. Using AI, a “smart” target discovery tool (SCMYSTERYDENTIFIER) was created to predict the immunosuppressive (IMMUSUP) function of mystery genes in endothelial cells (ECs), useful for alternative immunotherapy (ALTERIT) development. I will use my immunology/bioinformatics expertise to: (i) construct a single-cell vasculome atlas from publicly available data and characterize lung cancer ECs with immunomodulatory characteristics (IMECs) in-depth; (ii) use highly curated IMMUSUP gene sets to train the SCMYSTERYDENTIFIER tool to predict MYSTERY genes with IMMUSUP function; (iii) in vivo validate predicted IMMUSUP MYSTERY genes by generating EC-selective knockout mice rapidly (days)/inexpensively using a new in-house developed transgenic REVOLT technology. All approaches are operational. Perspectives: the developed AI-tool can be made generic to discover any type of function for mystery genes in any cell type.

Keywords:Endothelial cell, Target discovery, mystery genes

Disciplines:Single-cell data analysis, Cancer biology, Cellular interactions and extracellular matrix, Molecular and cell biology not elsewhere classified