Small molecules that target RNA are emerging as a powerful therapeutic modality, although deriving structure–activity relationships (SARs) remains a major challenge. Here, we present AI ‐augmented I terative S creening of L ibraries A gainst R NA targets (AISLAR), a machine learning‐driven strategy that accelerates the discovery of SAR‐tractable RNA binders and enables rational analog design. We screened diverse, drug‐like chemical libraries against two RNA motifs derived from human p53 mRNA and applied AISLAR within the open‐source KNIME platform. The application of AISLAR yielded chemotypes suitable for SAR development. Biophysical assays confirmed direct binding of representative compounds to one RNA motif. Guided by early SAR trends, we developed a pharmacophore hypothesis and designed an analog that retained binding with lower predicted cardiac channel liability. Docking simulations using the crystal structure of the RNA motif revealed a plausible binding mode for the validated hit compound. While further validation across diverse RNA targets and compound libraries will be required, these results demonstrate how AISLAR can be used as a workflow linking RNA‐targeted small molecule screening with rational analog design.
Hattori et al. (Wed,) studied this question.