A NOVEL REGRESSION-BASED ENSEMBLE CONVOLUTIONAL NEURAL NETWORK APPROACH FOR AUTOMATIC CANCER CELL DRUG SENSITIVITY PREDICTION

Abstract

Introduction: In modern medical advancements, the neural drug designing and sensitivity prognosis put a pace forward to novel methodologies and focused to reach the goal of predicting anti-cancer compound sensitivity by implementing multimodal-based convolution encoder. This novel approach is executed in three major key moves:  established expertise in the internal cellular dynamics derived based on protein interactions alongside tumor gene expression data, chemical structures represented with SMILES sequences. Our multi-scale convolutional attention-based encoder achieves an R2 value of 0.86 and an RMSE of 0.89, significantly surpassing a benchmark mechanism that integrates Morgan fingerprints, with numerous SMILES-based encoding strategies, and the established conventional methods for multi-modal drug sensitivity prognosis. In addition, the study proposes an Ensemble Convolution Neural Network Model: A Novel Regression-Based Approach (ECNN-NRNN) for pharma logical drug response assessment, which leverages numerous pharmacogenomic datasets while addressing the diversity in selected features for sub-pharma comic attributes. Given that certain pharma-genomic information is reachable to public in online mode, the vital attention to be done on therapeutic response and formulation.  The advancements in drug-sensitivity prognosis is able to attain through state-of-art mechanisms, and we will deliver empirical evaluation to demonstrate these improvements.