JoPMol Model Advances Precision Molecular Design with Joint Control

The study introduces JoPMol, a novel generative model designed for precision molecular design, aiming to discover personalized drug candidates. This model uniquely integrates various control conditions, including biological relevance derived from gene expression profiles, molecular structure information expressed in text, and quantified chemical properties. JoPMol's unified framework allows for the simultaneous generation and optimization of candidate molecules by jointly controlling these diverse conditions. Experimental evaluations demonstrate that JoPMol surpasses state-of-the-art methods across several metrics. Furthermore, the model exhibits strong generalization capabilities in transfer tasks and biologically realistic simulation scenarios, validating its effectiveness for advanced molecular design.