In silico designing of small molecules for targeting RNA: current landscape and future directions

RNA molecules govern nearly every layer of cellular regulation, yet they have long remained an underexploited target for small-molecule drug discovery relative to proteins. The 2020 FDA approval of risdiplam, the first small molecule to directly correct pre-mRNA splicing, demonstrated that folded RNAs can present druggable pockets and renewed interest in the field. However, fully exploiting the therapeutic potential of RNA depends on the development of computational methods capable of addressing (or utilizing or harnessing) the unique physicochemical characteristics of RNA It includes conformationally dynamic backbone, featureless binding surfaces, and the limited availability of high-resolution RNA-ligand co-crystal structures. This review surveys the computational pipeline built around these constraints, covering binding-site prediction (from early network-based tools to RNA language models such as RLsite and RNABind), RNA-specific docking and scoring functions (SPRank, RLDOCKScore, RNAmigos2), structure-based design (SILCS-RNA), fragment-based approaches guided by NMR screening, sequence-based platforms (Inforna), molecular dynamics and free-energy methods, and deep-learning affinity predictors (DeepRSMA, RLASIF, SMRTnet, PRISM). It concludes by examining the field's persistent constraints, including lack of structural data, underdeveloped force fields, unresolved selectivity, and the gap between in silico and cellular activity, and the advances likely to narrow them.