A New Paradigm for RNA Homology Search Empowered by Large Language Models and Contrastive Learning

Zihou Jiang

doi:10.54254/2755-2721/2026.TJ30347

Applied and Computational EngineeringOpen access

A New Paradigm for RNA Homology Search Empowered by Large Language Models and Contrastive Learning

Research Article

Open Access

A New Paradigm for RNA Homology Search Empowered by Large Language Models and Contrastive Learning

Zihou Jiang ^1*

¹ Shanghai Pinghe Bilingual School, Shanghai, China

^*Corresponding author: jiangzihou2020@163.com

Published on 3 December 2025

ACE Vol.211

ISSN (Print): 2755-273X

ISSN (Online): 2755-2721

ISBN (Print): 978-1-80590-579-0

ISBN (Online): 978-1-80590-580-6

Download Cover

Abstract

RNA’s growing therapeutic impact demands fast, structure-aware comparative analysis. Existing search tools trade speed for sensitivity: sequence-only methods are rapid but miss distant, structure-conserved homologs, whereas covariance-model pipelines are accurate but slow. We present a two-stage framework that reframes RNA homology detection as geometric retrieval in a learned embedding space, followed by structure-aware multiple sequence alignment (MSA). A frozen RNA foundation model (Rinalmo) embeds sequences; a lightweight Transformer head trained with supervised contrastive learning (family-balanced sampling, clan-aware hard negatives) sculpts the space so homologs cluster and non-homologs separate. Approximate nearest-neighbor search (FAISS/HNSW) enables sub-linear retrieval from millions of sequences. Top-khits are then aligned via a hybrid pipeline—MAFFT X-INS-i seeding and Infernal covariance-model refinement—to produce structure-consistent MSAs. On a family-level split of Rfam v14.10, Our method answers a query in 0.45 s on average (∼20×faster than BLASTn;>3,500×faster than cmscan) while achieving 0.95 precision, 0.93 recall, and 0.94 F1. Using retrieved sets, the MAFFT→Infernal workflow attains SPS = 0.91 versus 0.68 for BLASTn-based sets, enabling scalable, sensitive RNA homology discovery and downstream analysis.

Keywords:

RNA Homology Search, Deep Metric Learning, Multiple Sequence Alignment (MSA), Contrastive Learning

View PDF

References

[1]. Draper, D. E. A guide to ions and rna structure. RNA 10, 335–343 (2004). URL https: //rnajournal.cshlp.org/content/10/3/335.

[2]. Nawrocki, E. P. & Eddy, S. R. Infernal 1.1: 100-fold faster rna homology searches. Bioin- formatics 29, 2933–2935 (2013). URL https: //academic.oup.com/bioinformatics/ article/29/22/2933/316439.

[3]. Tommaso, P. D. et al. T-coffee: a web server for the multiple sequence alignment of protein and rna sequences using structural information and homology extension. Nucleic Acids Research 39, W13–W17 (2011). URL https: //pubmed.ncbi.nlm.nih.gov/21558174/.

[4]. Jumper, J. et al. Highly accurate protein structure prediction with alphafold. Nature 596, 583–589 (2021). URL https: //www.nature.com/articles/s41586-021-03819-2.

[5]. Chen, Y. et al. Hs-blastn: a platform-independent, high-speed nucleotide database search tool for the next-generation sequencing era. Nucleic Acids Research 43, 7762–7768 (2015). URL https: //academic.oup.com/nar/article/43/16/7762/1077466.

[6]. Zhang, T. et al. Rnacmap: a fully automatic pipeline for predicting contact maps of rnas by evolutionary coupling analysis. Bioinformatics 37, 3494–3500 (2021). URL https: //pubmed.ncbi.nlm.nih.gov/34021744/.

[7]. Eggenhofer, F., Hofacker, I. L. & zu Siederdissen, C. H. Rnalien: Unsupervised rna family model construction. Nucleic Acids Research 44, 8433–8441 (2016). URL https: //academic.oup.com/nar/article/44/17/8433/2468316.

[8]. Zhang, T., Singh, J. & Zhou, Y. rmsa: A sequence search and alignment algorithm to compute accurate rna homologs. Journal of Molecular Biology 435, 167969 (2023). URL https: //www.sciencedirect.com/science/article/pii/S0022283622005709.

[9]. Singh, J., Hanson, J., Paliwal, K. & Zhou, Y. Rna secondary structure prediction using an ensemble of two-dimensional deep neural networks and transfer learning. Nature Communi- cations 10, 5407 (2019). URL https: //www.nature.com/articles/s41467-019-13395-9.

[10]. Kalvari, I. et al. Rfam 14: expanded coverage of metagenomic, viral and microrna families. Nucleic acids research 49, D192–D200 (2021).

[11]. Penić, R. J., Vlašić, T., Huber, R. G., Wan, Y. & Šikić, M. Rinalmo: General-purpose rna language models can generalize well on structure prediction tasks. Nature Communications 16, 5671 (2025).

[12]. Khosla, P. et al. Supervised contrastive learning. Advances in neural information processing systems 33, 18661–18673 (2020).

[13]. Katoh, K. & Standley, D. M. Mafft multiple sequence alignment software version 7: im- provements in performance and usability. Molecular biology and evolution 30, 772–780 (2013).

[14]. Nawrocki, E. P. & Eddy, S. R. Infernal 1.1: 100-fold faster rna homology searches. Bioin- formatics 29, 2933–2935 (2013).

References

[1]. Draper, D. E. A guide to ions and rna structure. RNA 10, 335–343 (2004). URL https: //rnajournal.cshlp.org/content/10/3/335.

[4]. Jumper, J. et al. Highly accurate protein structure prediction with alphafold. Nature 596, 583–589 (2021). URL https: //www.nature.com/articles/s41586-021-03819-2.

[10]. Kalvari, I. et al. Rfam 14: expanded coverage of metagenomic, viral and microrna families. Nucleic acids research 49, D192–D200 (2021).

[12]. Khosla, P. et al. Supervised contrastive learning. Advances in neural information processing systems 33, 18661–18673 (2020).

[13]. Katoh, K. & Standley, D. M. Mafft multiple sequence alignment software version 7: im- provements in performance and usability. Molecular biology and evolution 30, 772–780 (2013).

[14]. Nawrocki, E. P. & Eddy, S. R. Infernal 1.1: 100-fold faster rna homology searches. Bioin- formatics 29, 2933–2935 (2013).

Cite this article

Jiang,Z. (2025). A New Paradigm for RNA Homology Search Empowered by Large Language Models and Contrastive Learning. Applied and Computational Engineering,211,27-41.

Data availability

The datasets used and/or analyzed during the current study will be available from the authors upon reasonable request.

About volume

Volume title: Proceedings of CONF-SPML 2026 Symposium: The 2nd Neural Computing and Applications Workshop 2025

ISBN: 978-1-80590-579-0(Print) / 978-1-80590-580-6(Online)

Editor: Marwan Omar, Guozheng Rao

Conference website: https://www.confspml.org/tianjin.html

Conference date: 21 December 2025

Series: Applied and Computational Engineering

Volume number: Vol.211

ISSN: 2755-2721(Print) / 2755-273X(Online)

© 2024 by the author(s). Licensee EWA Publishing, Oxford, UK. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license. Authors who publish this series agree to the following terms:

1. Authors retain copyright and grant the series right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this series.

2. Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the series's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this series.

3. Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See Open access policy for details).