A Balanced View Of Nucleic Acid Structural Modeling

This article collection provides a comprehensive overview of recent advances in RNA structure prediction and analysis, highlighting both computational and experimental approaches. The collection explores the growing application of deep learning techniques to RNA secondary structure prediction, addressing the limitations of traditional biophysics-based algorithms and the challenges imposed by training data biases. It introduces synthetic data as a means to assess machine learning model capabilities, and details robust deep learning algorithms, such as ATTfold, that leverage attention mechanisms to predict complex RNA structures—including long sequences and those with pseudoknots—with much greater accuracy than classical methods. Complementing these predictive techniques, the collection discusses tools like RNAStat, which enables thorough statistical analysis of RNA 3D structures and motifs, providing valuable resources for RNA modeling. Moreover, the integration of all-atom molecular dynamics simulations with experimental data, such as SAXS, is presented as a powerful approach to unravel the dynamic nature and conformational ensembles of RNA molecules, thereby enhancing biophysical understanding essential for gene function and therapeutic applications. Together, these contributions underscore the significant progress and future potential of computational and data-driven methods in elucidating RNA structure and function.
----
Nucleic acid modeling is a changing field. A new alphabet has arisen, complementing the Leontis-Westhof nomenclature, and which enables a more complete representation of nucleic acid structure. Recent success in solving protein structure by deep learning has turned RNA into the next frontier in structure prediction. The Cryo-Electron Microscopy revolution has created a demand for new fitting methods for nucleic acids. Under all of this lies a substrate of dynamical modeling and quality evaluation software.
In this Research Topic, we aim to bring together a balanced view of the multiple threads of  development in this vibrant and changing field. We wish to present a balanced view of current work in nucleic acid modeling, spanning alphabets, density map fitting, and structure prediction, as well as the underlying dynamics and quality control technology.
We solicit reviews and original research articles in the following areas:

Structural alphabets
Density map fitting
Structure prediction methods
Dynamics
Quality evaluation methods