Toward Efficient Drug Target Identification via Quantum-Enhanced Evolutionary Latent Modelling

Abstract

Drug discovery is a complicated and resource demanding procedure that entails the identification of a molecular candidate meeting biological activity, chemical and synthesis viability. The classical computational drug discovery methods have the challenge of the chemical space in the form of its large size and the complexity of molecular optimization in combinatorics. It has been suggested in this work that a quantum-enhanced version of target drug discovery can be developed by combining latent evolutionary optimization with synthesis-determined molecular prioritization. Generative models represent molecules in a continuous latent space and can be used to evolve molecules efficiently with pharmacological aims. The quantum optimization algorithms are used to speed up the fitness assessment and multi-objective selection. The strategy that is proposed provides the advantages that optimized molecules do not just have to be biologically effective, but they must also be chemically synthesizable, which enhances the translational viability between in-silico designs and laboratory synthesis.