Linking CD spectra to solution conformational dynamics in miRNA via an integrated experimental and computational approach

Understanding the structural dynamics of RNA is essential for deciphering its biological functions and advancing the design of RNA-based therapeutics. MicroRNAs (miRNAs), such as hsa-miR-145, present unique challenges due to their conformational heterogeneity and dynamic behavior in solution. Here, we introduce a data-driven approach combining synchrotron CD spectroscopy, enhanced sampling simulations, and computational CD spectra calculations to elucidate the structural dynamics of the hsa-miR-145 duplex and its single-stranded components. By leveraging thousands of computed spectra and structural conformers, we apply an unsupervised classification framework that systematically maps conformational states to their spectroscopic features. Experiments reveal that the duplex adopts an A-form RNA helical structure with a cooperative melting transition, while the single strands remain highly disordered. Clustering methods, based on spectral similarity and eRMSD metrics, identify structural motifs such as hybridization, mismatches, and non-canonical interactions, uncovering the relationships between RNA conformations and CD signal variations. This integration of experimental and computational methodologies highlights the power of data-driven strategies to analyze high-dimensional structural and spectral datasets, offering a pathway for understanding RNA dynamics and guiding the rational design of RNA-based therapeutics.