AI-guided prediction of ncRNA biochemical features for therapeutic targeting

Noncoding RNAs (ncRNAs) are increasingly recognized as important therapeutic targets across human diseases, driving advances in RNA-targeting modalities. However, successful therapeutic engagement depends on key biochemical properties of the target ncRNA, including subcellular localization, higher-order structure, and accessibility, which influence the suitability of different RNA-targeting modalities. Historically, resolving these properties relied on labor-intensive experimentation, limiting rational prioritization of RNA-targeting modalities. Emerging artificial intelligence (AI) tools can now predict ncRNA biochemical features at scale, enabling systematic characterization of target properties relevant to therapeutic design. In this opinion article, we outline AI approaches for predicting ncRNA localization, structure, and accessibility, including deep learning models, RNA structural predictors, and therapeutic prioritization frameworks. We discuss how these tools may support informed selection of RNA-targeting modalities, while highlighting current limitations and future opportunities for AI-guided therapeutic development.