AutoRNA: RNA tertiary structure prediction using variational autoencoder.
Kazanskii, M. A.; Uroshlev, L.; Zatylkin, F.; Pospelova, I.; Kantidze, O.; Gankin, Y.
Show abstract
Understanding the tertiary structure of RNA is essential for advancing therapeutic development and vaccine design. Traditional methods, such as dynamic simulations, have been employed to study RNA structure, but often struggle to capture the complex, non-linear relationships within RNA sequences. Many previous approaches have relied on simpler models, which limited their performance. The scarcity of data is a key challenge in predicting the tertiary structure of RNA. Therefore, herein we introduce a variational autoencoder (VAE), AutoRNA, that achieves a root mean square error (RMSE) of approximately 4.5 [A] in predicting nucleotide positions.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Concurrent prediction of RNA secondary structures with pseudoknots and local 3D motifs in an Integer Programming framework 96%
- Deep learning models for RNA secondary structure prediction (probably) do not generalise across families 95%
- DELPHI: accurate deep ensemble model for protein interaction sites prediction 95%
Similar papers in this journal
- sincFold: end-to-end learning of short- and long-range interactions in RNA secondary structure 98%
- A Reproducibility Analysis-based Statistical Framework for Residue-Residue Evolutionary Coupling Detection 96%
- RNAdvisor: a comprehensive benchmarking tool for the measure and prediction of RNA structural model quality 96%
Similar papers in this journal
- SpatialPPI: three-dimensional space protein-protein interaction prediction with AlphaFold Multimer 95%
- DeepNeuropePred: a robust and universal tool to predict cleavage sites from neuropeptide precursors by protein language model 94%
- DrugForm-DTA: Towards real-world drug-target binding Affinity Model 94%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.