Recent trends in RNA informatics: a review of machine learning and deep learning for RNA secondary structure prediction and RNA drug discovery

Abstract

Computational analysis of RNA sequences constitutes a crucial step in the field of RNA biology. As in other domains of the life sciences, the incorporation of artificial intelligence and machine learning techniques into RNA sequence analysis has gained significant traction in recent years. Historically, thermodynamics-based methods were widely employed for the prediction of RNA secondary structures; however, machine learning-based approaches have demonstrated remarkable advancements in recent years, enabling more accurate predictions. Consequently, the precision of sequence analysis pertaining to RNA secondary structures, such as RNA–protein interactions, has also been enhanced, making a substantial contribution to the field of RNA biology. Additionally, artificial intelligence and machine learning are also introducing technical innovations in the analysis of RNA–small molecule interactions for RNA-targeted drug discovery and in the design of RNA aptamers, where RNA serves as its own ligand. This review will highlight recent trends in the prediction of RNA secondary structure, RNA aptamers and RNA drug discovery using machine learning, deep learning and related technologies, and will also discuss potential future avenues in the field of RNA informatics.

Comparative Assessment of Pose Prediction Accuracy in RNA–Ligand Docking

“…Scoring functions may need to reflect this . Also, ligand binding sites on RNA can be less deep and more polar, solvated, and conformationally flexible than for proteins, suggesting a possible need for modification of the search methods. − Apart from these, the electrostatic parameters could be updated to capture aromatic ring-stacking interactions, charged phosphate group interactions, and the participation of metal ions such as Mg 2+ and water molecules, which are critical for ligand–RNA interactions. ,,, Furthermore, small-scale conformational dynamics in the pocket are important for binding, so methods such as ensemble or multiconformer docking methods may need to be used to capture the dynamic nature of RNA. ,− Further, an alternative approach pursued in protein–ligand complexes involves the use of machine learning (ML) models to predict interactions, binding affinities, and ML-based scoring functions. ,,, The effectiveness of machine learning methods such as support vector machines (SVM), random forests (RF), neural networks (NN), and convolutional neural networks (CNN) has been demonstrated in protein–ligand systems. − These advances have the potential to enhance predictions of RNA–small molecule interactions using ML methods. For example, RNAPosers, which utilizes a random forest method, estimates ligand “nativeness” within an RNA–ligand structure.…”

“…51,56−58 Further, an alternative approach pursued in protein−ligand complexes involves the use of machine learning (ML) models to predict interactions, binding affinities, and ML-based scoring functions. 42,43,46,59 The effectiveness of machine learning methods such as support vector machines (SVM), random forests (RF), neural networks (NN), and convolutional neural networks (CNN) has been demonstrated in protein−ligand systems. 60−65 These advances have the potential to enhance predictions of RNA−small molecule interactions using ML methods.…”

Application of artificial scaffold systems in microbial metabolic engineering

“…However, its discrete structure prevents the formation of complex geometric shapes in living organisms, limiting its application compared to other types of scaffolds ( Lv et al, 2020 ). AI models also open up new opportunities for predicting the secondary and tertiary structures of RNA molecules, providing insights into designing new functions and interactions and improving stability ( Sato and Hamada, 2023 ).…”

Codon-optimization in gene therapy: promises, prospects and challenges

“…For example, to date (February 2024) only 6,142, of which 1,416 are human, experimentally validated RNA structures have been deposited in the Protein Data Bank ( Berman, 2000 ). This indicates that the high-precision prediction of RNA 3D structures using machine learning methods may be accurate for training data, but not for new data ( Sato and Hamada, 2023 ).…”