Development of targeted combination therapy and novel tools to predict therapy resistance

Therapeutic resistance persists as a cardinal impediment in oncology, with onset possible both pre- and intra-treatment. Consequently, predictive tools discerning drug efficacy can forestall initial therapeutic inadequacies and facilitate the selection of suitable alternative drugs post-resistance. At present, such predictive capabilities are deficient. Our recent findings indicate that transcriptional perturbations during drug treatment provide valuable data for the construction of machine-learning models for drug sensitivity prediction. In a preliminary application, we employed a deep learning-based approach to create a venetoclax sensitivity prediction model, further elucidating venetoclax resistance's molecular mechanisms. The gathered data suggest a putative interplay between lipid metabolism and polo-like kinase 1 (PLK1) signaling to facilitate venetoclax resistance, influencing the transcriptional regulation of apoptosis master regulators, specifically PMAIP1 and BCL2L13 within the B-cell lymphoma 2 (BCL2) gene family. Nonetheless, the specifics of lipid metabolism's interaction with PLK1, and the viability of its concurrent inhibition with PLK1 signaling for therapeutic combinations, warrant further investigation. This project aims to develop advanced therapeutic response predictors and elucidate the lipid metabolism-PLK1 signaling interaction in therapy resistance, prioritizing the identification of drug combinations to surmount such resistance.