Comparative Assessment of Pose Prediction Accuracy in RNA-Ligand Docking

SPRank─A Knowledge-Based Scoring Function for RNA-Ligand Pose Prediction and Virtual Screening

The growing interest in RNA-targeted drugs underscores the need for computational modeling of interactions between RNA molecules and small compounds. Having a reliable scoring function for RNA-ligand interactions is essential for effective computational drug screening. An ideal scoring function should not only predict the native pose for ligand binding but also rank the affinity of the binding for different ligands. However, existing scoring functions are primarily designed to predict the native binding modes for a given RNA-ligand pair and have not been thoroughly assessed for virtual screening purposes. In this paper, we introduce SPRank, a combination of machine-learning and knowledge-based scoring functions developed through a weighted iterative approach, specifically designed to tackle both binding mode prediction and virtual screening challenges. Our approach incorporates third-party docking software, such as rDock and AutoDock Vina, to sample flexible ligands against an ensemble of RNA structures, capturing the conformational flexibility of both the RNA and the ligand. Through rigorous testing, SPRank demonstrates improved performance compared to the tested scoring functions across four test sets comprising 122, 42, 55, and 71 nucleic acid-ligand complexes. Furthermore, SPRank exhibits improved performance in virtual screening tests targeting the HIV-1 TAR ensemble, which highlights its advantage in drug discovery. These results underscore the advantages of SPRank as a potentially promising tool for the RNA-targeted drug design. The source code of SPRank and the data sets are freely accessible at https://github.com/Vfold-RNA/SPRank.

Can Current Molecular Docking Methods Accurately Predict RNA Inhibitors?

Ribonucleic acids (RNAs), particularly the noncoding RNAs, play key roles in cancer, making them attractive drug targets. While conventional methods such as high throughput screening are resource-intensive, computational methods such as RNA-ligand docking can be used as an alternative. However, currently available docking methods are fine-tuned to perform protein-ligand and protein-protein docking. In this work, we evaluated three commonly used docking methods─AutoDock Vina, HADDOCK, and HDOCK─alongside RLDOCK, which is specifically designed for RNA-ligand docking. Our evaluation was based on several criteria including cognate docking, blind docking, scoring potential, and ranking potential. In cognate docking, only RLDOCK showed a success rate of 70% for the top-scoring docked pose. Despite this, all four docking methods did not achieve an overall success rate exceeding 50% amidst our attempt to refine the top-scoring docked poses using molecular dynamics simulations. Meanwhile, all four docking methods showed poor performance in scoring potential evaluation. Although AutoDock Vina achieved an area under the receiver operating characteristic curve of 0.70, it showed poor performance in terms of Matthews' correlation coefficient, precision, enrichment factors, and normalized enrichment factors at 1, 2, and 5%. These results highlight the growing need for further optimization of docking methods to assess RNA-ligand interactions.

Geometric deep learning for the prediction of magnesium-binding sites in RNA structures

Magnesium ions (Mg2+) are essential for the folding, functional expression, and structural stability of RNA molecules. However, predicting Mg2+-binding sites in RNA molecules based solely on RNA structures is still challenging. The molecular surface, characterized by a continuous shape with geometric and chemical properties, is important for RNA modelling and carries essential information for understanding the interactions between RNAs and Mg2+ ions. Here, we propose an approach named RNA-magnesium ion surface interaction fingerprinting (RMSIF), a geometric deep learning-based conceptual framework to predict magnesium ion binding sites in RNA structures. To evaluate the performance of RMSIF, we systematically enumerated decoy Mg2+ ions across a full-space grid within the range of 2 to 10 Å from the RNA molecule and made predictions accordingly. Visualization techniques were used to validate the prediction results and calculate success rates. Comparative assessments against state-of-the-art methods like MetalionRNA, MgNet, and Metal3DRNA revealed that RMSIF achieved superior success rates and accuracy in predicting Mg2+-binding sites. Additionally, in terms of the spatial distribution of Mg2+ ions within the RNA structures, a majority were situated in the deep grooves, while a minority occupied the shallow grooves. Collectively, the conceptual framework developed in this study holds promise for advancing insights into drug design, RNA co-transcriptional folding, and structure prediction.