Skip to content
DraconiticResearch
Search Sources API
182,460 studies31 sourcesUpdated 37 min ago
Peer-reviewedCC BY-NC 4.0RNA

Improving RNA secondary structure prediction through expanded training data

  • Conner J. Langeberg1,2,3
  • Taehan Kim4
  • Roma Nagle3,4
  • Agni Rajinikanth4
  • Charlotte Meredith3
  • Dimple Amitha Garuadapuri5,6
  • Jennifer A. Doudna1,2,3,7,8,9
  • Jamie H.D. Cate1,3,8,9
9 affiliations
  1. 1Innovative Genomics Institute, University of California, Berkeley, California 94720, USA
  2. 2California Institute for Quantitative Biosciences (QB3), University of California, Berkeley, California 94720, USA
  3. 3Department of Molecular and Cell Biology, University of California, Berkeley, Berkeley, California 94720, USA
  4. 4Department of Computer Science, University of California, Berkeley, Berkeley, California 94720, USA
  5. 5Department of Bioengineering, University of California, Berkeley, Berkeley, California 94720, USA
  6. 6Department of Electrical Engineering and Computer Science, University of California, Berkeley, Berkeley, California 94720, USA
  7. 7Howard Hughes Medical Institute, University of California, Berkeley, Berkeley, California 94720, USA
  8. 8Department of Chemistry, University of California, Berkeley, Berkeley, California 94720, USA
  9. 9MBIB Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, USA
Read full textPDFDOICite

Abstract

In recent years, deep learning has revolutionized protein structure prediction, achieving remarkable speed and accuracy. RNA structure prediction, however, has lagged behind. Although several methods have shown moderate success in predicting RNA secondary and tertiary structures, none have reached the accuracy observed with contemporary protein models. The lack of success of these RNA structure prediction models has been proposed to be due to limited high-quality structural information that can be used as training data. To probe this proposed limitation, we developed a large and diverse data set comprising paired RNA sequences and their corresponding secondary structures. We assessed the utility of this enhanced data set by retraining a deep-learning model, SincFold. We find that SincFold exhibited improved performance on a set of previously unseen RNA families, enhancing its capability to predict accurate de novo RNA secondary structures. We additionally implemented Lyra-TransPred, which achieved the highest mean F1 and MCC among the evaluated models while requiring substantially less training time per epoch. The RNASSTR data set provides a substantial advance for RNA structure modeling, laying a strong foundation for the development of future RNA secondary structure prediction algorithms.

Related

  1. CBAT-Net: a unified CNN–BiLSTM–MultiHead transformer architecture for joint Q3/Q8 protein secondary structure prediction
    Ridwan Uz Zaman Pandit, Jagjit Singh Dhatterwal, Gyana Ranjana Panigrahi +1
    Scientific ReportsPeer-reviewedCC BY-NC-ND 4.0Full textMolecular biology & gene regulationCrossref
  2. Leveraging secondary-structure information for accurate nucleic acid structure prediction with OFoldNA
    Valhalla Team
    arXivPreprintCC BY-NC-SA 4.0Full textBiophysicsarXiv
  3. A Graphormer approach for predicting peptide half-life, secondary structure, solvent accessibility, and post-translational modifications
    Paula Cifuentes, Ramon Adàlia, Lisa A. Vasicek +3
    Scientific ReportsPeer-reviewedCC BY-NC-ND 4.0Full textMolecular biology & gene regulationCrossref
  4. Tubulin E-Hook Hexamers Reveal Sequence-Dependent Compaction and Transient Secondary Structure Signatures.
    Alexander C Bromley, Nicholas A Kruse, Connor R Brower +4
    Molecules (Basel, Switzerland)Peer-reviewedCC BYFull textMolecular biology & gene regulationEurope PMC
  5. MERGE-RNA: a physics-based model to predict RNA secondary structure ensembles with chemical probing
    Giuseppe Sacco, Jianhui Li, Redmond Smyth +2
    Nucleic Acids ResearchPeer-reviewedCC BYMolecular biology & gene regulationHAL

Cite

BibTeXRIS
BibTeX
@article{langeberg2026improving,
  title = {{Improving RNA secondary structure prediction through expanded training data}},
  author = {Conner J. Langeberg and Taehan Kim and Roma Nagle and Agni Rajinikanth and Charlotte Meredith and Dimple Amitha Garuadapuri and Jennifer A. Doudna and Jamie H.D. Cate},
  journal = {RNA},
  year = {2026},
  month = nov,
  doi = {10.1261/rna.081259.126},
  url = {https://doi.org/10.1261/rna.081259.126},
  note = {CC BY-NC 4.0},
}
RIS
TY  - JOUR
TI  - Improving RNA secondary structure prediction through expanded training data
AU  - Conner J. Langeberg
AU  - Taehan Kim
AU  - Roma Nagle
AU  - Agni Rajinikanth
AU  - Charlotte Meredith
AU  - Dimple Amitha Garuadapuri
AU  - Jennifer A. Doudna
AU  - Jamie H.D. Cate
PY  - 2026
DA  - 2026/11/01
JO  - RNA
DO  - 10.1261/rna.081259.126
UR  - https://doi.org/10.1261/rna.081259.126
AB  - In recent years, deep learning has revolutionized protein structure prediction, achieving remarkable speed and accuracy. RNA structure prediction, however, has lagged behind. Although several methods have shown moderate success in predicting RNA secondary and tertiary structures, none have reached the accuracy observed with contemporary protein models. The lack of success of these RNA structure prediction models has been proposed to be due to limited high-quality structural information that can be used as training data. To probe this proposed limitation, we developed a large and diverse data set comprising paired RNA sequences and their corresponding secondary structures. We assessed the utility of this enhanced data set by retraining a deep-learning model, SincFold. We find that SincFold exhibited improved performance on a set of previously unseen RNA families, enhancing its capability to predict accurate de novo RNA secondary structures. We additionally implemented Lyra-TransPred, which achieved the highest mean F1 and MCC among the evaluated models while requiring substantially less training time per epoch. The RNASSTR data set provides a substantial advance for RNA structure modeling, laying a strong foundation for the development of future RNA secondary structure prediction algorithms.
SN  - 1355-8382
SN  - 1469-9001
ER  -