Key points are not available for this paper at this time.
Identifying precise molecular targets is hindered by intricate disease mechanisms and the necessity for specificity. Biological noise and emerging resistance mechanisms demand personalized and combination therapies. Artificial intelligence (AI) holds promise for improving drug design by analyzing large omics datasets to pinpoint relevant targets, reducing drug candidate failures, and expediting preclinical testing through in silico methods. However, challenges include biological complexity, feature selection, data quality, and model interpretability. Despite advances in AI for drug screening and target identification, enhancing molecular-level drug responses is a significant challenge. Biological noise, gene expression fluctuations, and cellular dynamics are vital considerations for bioengineering. This paper tackles current drug development challenges, emphasizing the role of biological noise at the molecular level. We describe noise-related challenges in structure-based drug design and highlight the need to account for protein and ligand flexibility in molecular docking and molecular dynamics simulations. Molecular mechanics and genetic algorithms are also explored as strategies to optimize drug binding. The constrained disorder principle (CDP)-based second-generation AI platform is being developed for drug development. The paper provides a perspective on how the CDP-based AI platform can address challenges at the pre-target, target, and post-target levels. It suggests techniques to incorporate variability into drug design and administration, ultimately leading to more effective and personalized therapeutic strategies.
Hurvits et al. (Tue,) studied this question.