Pith. sign in

REVIEW 4 major objections 6 minor 72 references

A Comparative Review of RNA Language Models

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read RNA language models face a structure-versus-function trade-off.

desk verdict A useful benchmark of RNA language models with a plausible but under-supported trade-off claim; the OR metric needs validation before the headline conclusion is accepted. read the letter →

arxiv 2505.09087 v1 pith:KHQNJXA2 submitted 2025-05-14 q-bio.BM cs.LG

He Wang , Yikun Zhang , Jie Chen , Jian Zhan , Yaoqi Zhou This is my paper · ORCID
classification q-bio.BMcs.LG
keywords RNAlanguagemodelszero-shotpredictionsecondarystructureclassificationoverlapratioembeddingsimilaritypretrainingobjectivesnoncoding
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper compares 13 RNA language models plus DNA and protein controls under a common zero-shot evaluation, testing each model's ability to predict RNA secondary structure from attention maps and to distinguish homologous from non-homologous RNA sequences using learned embeddings. The central finding is that most models excel at one task at the expense of the other: models with strong secondary-structure prediction, such as RNA-km and MP-RNA, fare poorly at functional classification, while models with strong functional classification, such as RNA-FM, lag in structure. Only AIDO.RNA achieves high performance on both axes. The authors conclude that current pretraining objectives are unbalanced and that RNA language models need training schemes that improve structure prediction without degrading functional representation.

What carries the argument

Two zero-shot probes carry the comparison. For structure, the paper extracts per-head attention maps from each transformer, symmetrizes and applies APC correction, then thresholds the resulting probabilities into base-pair predictions, selecting the optimal head-layer and threshold on a 40-RNA validation set and scoring on a 70-RNA test set. For function, the paper computes the overlap ratio (OR), defined as the intersection area of the cosine-similarity distributions of homologous versus non-homologous sequence pairs divided by the area of their union; OR ranges from 0 (perfectly separated distributions) to 1 (fully overlapping), so a smaller OR means better functional discrimination. The FFT-compressed sequence embeddings used for cosine similarity are the key methodological choice that makes OR computable for all models regardless of embedding dimension.

What would settle it

Take the same set of RNA language models and evaluate them on a supervised RNA family classification task, such as training a linear classifier on their sequence embeddings for Rfam families, then compare ranks with OR ranks. If models with low OR do not systematically achieve higher supervised classification accuracy, or if a model with high OR clearly outperforms one with low OR, the central trade-off claim would be falsified. A simpler check: recompute OR using mean-pooled embeddings instead of FFT-compressed vectors and see whether the pattern of model rankings persists.

Watch

Extended reading notes

Core claim

Across a standardized zero-shot comparison, RNA language models show an inverse relationship between secondary-structure prediction quality and embedding-based functional classification quality. Secondary-structure F1 was measured by converting attention maps from each model into base-pair probabilities and scoring against experimentally determined RNA structures, while functional classification was measured by a newly introduced overlap ratio (OR) between cosine-similarity distributions of homologous and non-homologous sequence pairs. Models that rank highest in secondary-structure F1, including RNA-km, MP-RNA, and RNA-MSM, generally show worse OR values, whereas models like RNA-FM and ERNIE-RNA with low OR values show weaker structure prediction. AIDO.RNA (1.6B) is the only model with both strong secondary-structure F1 and low OR, although increasing its size from 650M to 1.6B improved structure but worsened classification. The paper interprets this as evidence that current unsupervised pretraining produces representations that are skewed toward either structural or functional signal, and argues for rebalanced training objectives.

Load-bearing premise

The overlap ratio computed from FFT-compressed embedding cosine similarities is a valid zero-shot measure of an RNA language model's functional classification ability, even though this metric has not been validated against supervised functional benchmarks.

Editorial extensions

If this is right

  • If the trade-off is real, future RNA language models should be evaluated on both structure and function simultaneously, using a shared protocol like this one, rather than on whichever downstream task each model was tuned to.
  • Pretraining objectives need rebalancing: strategies that inject structural bias, such as k-mer masking in RNA-km or secondary-structure calibration in MP-RNA, appear to degrade the model's ability to separate RNA families in embedding space.
  • Scaling model size alone is not a reliable path to balanced performance, since the larger AIDO.RNA improved secondary structure but worsened functional classification relative to the smaller variant.
  • MSA-based models such as RNA-MSM achieve the best secondary-structure F1 despite being trained on far fewer sequences, suggesting that evolutionary information is a strong complementary signal for structure that single-sequence pretraining does not capture.
  • Unified RNA-DNA-protein models, represented by LucaOne, fall in the mid-range for both tasks, indicating that a universal biological language model may sacrifice RNA-specific competence.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The OR metric is introduced here without external validation against supervised functional classification benchmarks; if OR does not track true functional relatedness, the claimed trade-off could be an artifact of embedding geometry rather than a real limitation of RNA language models.
  • A direct test would be to compare OR rankings with fine-tuned classification accuracy on the same Rfam and ArchiveII splits; a strong negative correlation between OR and supervised accuracy would confirm the metric, while disagreement would weaken the paper's central conclusion.
  • The FFT compression to 128 dimensions may discard information differently across models of different embedding sizes, potentially biasing OR comparisons; using mean-pooled embeddings or a fixed projection trained on one model could serve as a sensitivity check.
  • If the trade-off is confirmed, a practical next step is multi-task pretraining that jointly optimizes masked language modeling and a secondary-structure auxiliary objective, then measuring whether both OR and F1 improve together.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper benchmarks 13 RNA language models, three DNA LMs, and one protein LM under a zero-shot protocol. For secondary structure, it extracts attention maps from each model and computes F1 against experimental structures using threshold/head positions chosen on a small validation set. For functional classification, it introduces a new metric, the overlap ratio (OR), which measures the overlap between the cosine-similarity distributions of homologous and non-homologous sequence pairs sampled from Rfam and ArchiveII. The authors report that models strong on secondary structure often have poorer OR values, that AIDO.RNA (1.6B) is the only model with high performance on both tasks, and they conclude that current pretraining objectives are unbalanced.

Significance. If the central claim were supported, this would be a useful comparative resource for the RNA language-model community and would motivate rebalanced pretraining objectives. The paper's strengths include the breadth of models compared, the explicit zero-shot protocol before fine-tuning, the inclusion of DNA and protein controls, the documentation of rRNA/tRNA imbalance in common training and benchmark datasets, and the promise of open data and code on Zenodo. However, the central trade-off conclusion rests entirely on the newly introduced OR metric, which is not externally validated as a measure of functional classification. Because the functional half of the benchmark is load-bearing, the reported trade-off and the recommendation to rebalance pretraining are not yet established.

major comments (4)
  1. [Methods, 'Calculation of the overlap ratio (OR)'; Table 4] The OR metric is a distributional overlap between pairwise embedding similarities, not a measure of classification accuracy. It is introduced in this paper without validating it against any standard functional-classification benchmark, a baseline classifier, an AUC value, or a known functional category annotation. Since the central trade-off claim is built entirely on this metric, the authors should validate OR against a genuine supervised classification task (e.g., held-out Rfam family classification or RNA-type classification) and against simple baselines such as sequence identity or k-mer frequency. They should also report the variance of OR across models and runs; the Table 4 entries of 0.000 or 0.001 give the impression of zero variability and do not support the rank comparisons made in the text.
  2. [Results, 'Zero-shot model evaluation: RNA classifications'; Table 4] The claim that 'models doing well on secondary structure prediction often perform worse in function classification or vice versa' is supported only by selected examples and visual rank comparison. No correlation coefficient, significance test, or confidence interval is reported between the SS F1 values and OR values across models. The comparison is also incomplete because RNA-MSM and ESM2 are absent from the OR columns and several models are absent from the SS column. The authors should compute a quantitative correlation (with uncertainty), include all models that can be evaluated, and address obvious confounders such as model size and pretraining data composition before asserting a trade-off.
  3. [Results, 'Zero-shot model evaluation: RNA classifications'] The text states that Rfam families are easier to separate than ArchiveII families because sequence identity within Rfam families averages 77.1% versus 54.5% in ArchiveII. This admits that the OR values are likely driven by sequence identity and pretraining-corpus overlap rather than by functional knowledge. The paper should control for sequence identity, for example by stratifying homologous pairs into identity bins or by comparing OR with an alignment-free identity baseline. Without such a control, the interpretation of OR as a measure of functional classification ability is not justified.
  4. [Introduction, discussion of data redundancy; Methods, dataset preparation] The paper criticizes existing benchmarks for not performing redundancy removal, but its own zero-shot classification test uses Rfam and ArchiveII sequences that may overlap substantially with the RNAcentral or other corpora on which most of the evaluated models were pretrained. This potential leakage could inflate the apparent discrimination ability of the models. The authors should assess leakage directly, for example by removing training-set sequences that are nearly identical to test sequences, or by reporting how OR changes under different redundancy thresholds.
minor comments (6)
  1. [Abstract] The sentence 'Results shows that...' should be 'Results show that...'.
  2. [Table 1] The column header 'Art.' is unexplained; the authors should either define it or remove it.
  3. [References] The reference list contains duplicate numbering (two entries labeled 32); all subsequent reference numbers need renumbering.
  4. [Code Availability] The model name is misspelled as 'DNlABERT-2'; it should be 'DNABERT-2'.
  5. [Table 4] The meaning of the asterisk entries should be explained in more detail; for example, readers may want to know why DNABERT-2 and NT have no SS F1 values and why RNA-MSM and ESM2 have no OR values.
  6. [Extended Data Figs. 1 and 2] The axes and legend labels in the cosine-similarity distribution figures are not described in the caption; adding explicit axis titles and a legend for the homologous and non-homologous distributions would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the paper is an empirical cross-model benchmark with held-out evaluation; the trade-off claim is an observed pattern, not a derivation from its own inputs.

full rationale

This is a benchmark/comparison paper rather than a derivation chain. The central claim—that RNA language models trade off zero-shot secondary-structure performance against functional classification—is an empirical observation from Table 4 and Fig. 2, not a quantity derived from an assumed conclusion. The secondary-structure protocol selects head-layer positions and thresholds on the VL1 validation set and then evaluates on the held-out TS set, so the reported F1 scores are not fitted to the test set. The overlap ratio (OR) is a newly introduced metric, but it is computed directly from the cosine-similarity distributions of homologous and non-homologous sequence pairs; it is not fitted to the trade-off result and it does not presuppose which models will rank well. Whether OR is a valid proxy for functional classification is an important external-validity concern, but that is a correctness risk, not circularity. The paper does cite the authors' own RNA-MSM work for the attention-map extraction protocol and the VL1/TS datasets, and RNA-MSM is one of the models compared; however, this self-citation is not load-bearing in the sense of forcing the central conclusion. The same protocol is applied uniformly to all 13 RNA LMs and controls, and the comparative ranking is independent of any single model's own performance claims. No equation in the paper reduces to another by construction, no fitted parameter is relabeled as a prediction, and no uniqueness theorem or prior result is invoked to rule out alternatives. The central trade-off is an observed pattern from a held-out comparison, so the paper is self-contained as an empirical benchmark and no significant circularity is present.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new mathematical derivations, so the main free parameters are the validation-set threshold for structure and the hand-set embedding dimension for classification. The key axioms are the validity of attention-map-based structure extraction and the validity of the cosine-similarity OR metric for function.

free parameters (2)
  • SS attention threshold = selected per model on VL1, range [0, 1] at 0.001 steps
    The reported secondary-structure F1 depends on a threshold chosen on the VL1 validation set for each model; this is a fitted hyperparameter, though not fitted to the TS test set.
  • FFT embedding dimension d' = 128
    The classification metric compresses each sequence embedding to 128 dimensions with a Fourier transform before computing cosine similarity; the paper gives no justification for this dimension.
assumptions (3)
  • domain assumption Attention maps of BERT-style LMs encode RNA secondary structure after symmetrization, APC correction, and sigmoid thresholding.
    This is the basis of the zero-shot secondary structure protocol, originally from the authors' RNA-MSM paper (ref 20), and it is assumed to hold across all models.
  • ad hoc to paper Cosine similarity between FFT-compressed sequence embeddings is a valid proxy for RNA functional relatedness.
    The OR metric is introduced in this paper and has no external validation; the central trade-off claim depends on it.
  • domain assumption Rfam family labels and ArchiveII RNA-type labels are correct ground-truth classes for homology.
    The classification evaluation treats same-family (or same-type) sequences as homologous; this assumes dataset annotations are accurate.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Comparative Review of RNA Language Models." pith.science (2026). https://pith.science/paper/KHQNJXA2

@misc{pith2026250509087,
  author       = {Pith},
  title        = {Pith review of: A Comparative Review of RNA Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KHQNJXA2}},
  note         = {Machine review of arXiv:2505.09087}
}
read the original abstract

Given usefulness of protein language models (LMs) in structure and functional inference, RNA LMs have received increased attentions in the last few years. However, these RNA models are often not compared against the same standard. Here, we divided RNA LMs into three classes (pretrained on multiple RNA types (especially noncoding RNAs), specific-purpose RNAs, and LMs that unify RNA with DNA or proteins or both) and compared 13 RNA LMs along with 3 DNA and 1 protein LMs as controls in zero-shot prediction of RNA secondary structure and functional classification. Results shows that the models doing well on secondary structure prediction often perform worse in function classification or vice versa, suggesting that more balanced unsupervised training is needed.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

72 extracted references · 61 canonical work pages

  1. [1]

    Zhang, Q. et al. Scientific Large Language Models: A Survey on Biological & Chemical Domains. Preprint at http://arxiv.org/abs/2401.14656 (2024)

    arXiv 2024

  2. [2]

    & Toutanova, K

    Devlin, J., Chang, M.-W., Lee, K. & Toutanova, K. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. ArXiv E-Prints arXiv:1810.04805 (2018) doi:10.48550/arXiv.1810.04805

    arXiv doi:10.48550/arxiv.1810.04805 2018

  3. [3]

    & Narasimhan, K

    Radford, A. & Narasimhan, K. Improving Language Understanding by Generative Pre-Training. in (2018)

    work page 2018

  4. [4]

    Brown, T. B. et al. Language Models are Few-Shot Learners. ArXiv E-Prints arXiv:2005.14165 (2020) doi:10.48550/arXiv.2005.14165

    arXiv doi:10.48550/arxiv.2005.14165 2020

  5. [5]

    Touvron, H. et al. LLaMA: Open and Efficient Foundation Language Models. ArXiv E-Prints arXiv:2302.13971 (2023) doi:10.48550/arXiv.2302.13971

    arXiv doi:10.48550/arxiv.2302.13971 2023

  6. [6]

    Chowdhery, A. et al. PaLM: scaling language modeling with pathways. J Mach Learn Res 24, (2024)

    work page 2024

  7. [7]

    Raffel, C. et al. Exploring the Limits of Transfer Learning with a Unified Text-to- Text Transformer. J. Mach. Learn. Res. 21, 1–67 (2020)

    work page 2020

  8. [8]

    Vaswani, A. et al. Attention Is All You Need. ArXiv E-Prints arXiv:1706.03762 (2017) doi:10.48550/arXiv.1706.03762

    arXiv doi:10.48550/arxiv.1706.03762 2017

Show all 72 references
  1. [9]

    Elnaggar, A. et al. ProtTrans: Toward Understanding the Language of Life Through Self-Supervised Learning. IEEE Trans. Pattern Anal. Mach. Intell. 44, 7112– 7127 (2022)

    2022

  2. [10]

    & Höcker, B

    Ferruz, N., Schmidt, S. & Höcker, B. ProtGPT2 is a deep unsupervised language model for protein design. Nat. Commun. 13, 4348 (2022)

    2022

  3. [11]

    & Linial, M

    Brandes, N., Ofer, D., Peleg, Y., Rappoport, N. & Linial, M. ProteinBERT: a universal deep-learning model of protein sequence and function. Bioinforma. Oxf. Engl. 38, 2102–2110 (2022)

    2022

  4. [12]

    Rives, A. et al. Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences. Proc. Natl. Acad. Sci. U. S. A. 118, (2021)

    2021

  5. [13]

    Rao, R. et al. MSA Transformer. bioRxiv (2021) doi:10.1101/2021.02.12.430858

    2021 doi

  6. [14]

    Lin, Z. et al. Evolutionary-scale prediction of atomic-level protein structure with a language model. Science 379, 1123–1130 (2023)

    2023

  7. [15]

    Rao, R. et al. Evaluating Protein Transfer Learning with TAPE. Preprint at https://doi.org/10.1101/676825 (2019)

    2019 doi

  8. [16]

    & Davuluri, R

    Ji, Y., Zhou, Z., Liu, H. & Davuluri, R. V. DNABERT: pre-trained Bidirectional Encoder Representations from Transformers model for DNA-language in genome. Bioinforma. Oxf. Engl. 37, 2112–2120 (2021)

    2021

  9. [17]

    Chen, J. et al. Interpretable RNA Foundation Model from Unannotated Data for Highly Accurate RNA Structure and Function Predictions. (2022)

    2022

  10. [18]

    Wang, X. et al. UNI-RNA: UNIVERSAL PRE-TRAINED MODELS REVOLUTIONIZE RNA RESEARCH. Preprint at https://doi.org/10.1101/2023.07.11.548588 (2023)

    2023 doi

  11. [19]

    Yin, W. et al. ERNIE-RNA: An RNA Language Model with Structure-enhanced Representations. Preprint at https://doi.org/10.1101/2024.03.17.585376 (2024)

    2024 doi

  12. [20]

    Zhang, Y. et al. Multiple sequence alignment-based RNA language model and its application to structural inference. Nucleic Acids Res. 52, e3–e3 (2024)

    2024

  13. [21]

    Chu, Y. et al. A 5′ UTR language model for decoding untranslated regions of mRNA and function predictions. Nat. Mach. Intell. 6, 449–460 (2024)

    2024

  14. [22]

    He, Y. et al. LucaOne: Generalized Biological Foundation Model with Unified Nucleic Acid and Protein Language. Preprint at https://doi.org/10.1101/2024.05.10.592927 (2024)

    2024 doi

  15. [23]

    Zhang, R., Ma, B., Xu, G. & Ma, J. ProtRNA: A Protein-derived RNA Language Model by Cross-Modality Transfer Learning. Preprint at https://doi.org/10.1101/2024.09.10.612218 (2024)

    2024 doi

  16. [24]

    J., Vlašić, T., Huber, R

    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. Preprint at http://arxiv.org/abs/2403.00043 (2024)

    2024 arXiv

  17. [25]

    Wang, N. et al. Multi-purpose RNA language modelling with motif-aware pretraining and type-guided fine-tuning. Nat. Mach. Intell. 6, 548–557 (2024)

    2024

  18. [26]

    Gong, T. & Bu, D. Language models enable zero-shot prediction of RNA secondary structure including pseudoknots

  19. [27]

    Chen, K. et al. Self-supervised learning on millions of primary RNA sequences from 72 vertebrates improves sequence-based RNA splicing prediction. Brief. Bioinform. 25, bbae163 (2024)

    2024

  20. [28]

    Nguyen, E. et al. Sequence modeling and design from molecular to genome scale with Evo. Science 386, eado9336 (2024)

    2024

  21. [29]

    Nucleic Acids Res

    RNAcentral 2021: secondary structure integration, improved sequence search and new member databases. Nucleic Acids Res. 49, D212–D220 (2021)

    2021

  22. [30]

    Yang, H. & Li, K. MP-RNA: Unleashing Multi-species RNA Foundation Model via Calibrated Secondary Structure Prediction

  23. [31]

    & Zhou, Y

    Chen, K., Litfin, T., Singh, J., Zhan, J. & Zhou, Y. MARS and RNAcmap3: The Master Database of All Possible RNA Sequences Integrated with RNAcmap for RNA Homology Search. Genomics Proteomics Bioinformatics 22, qzae018 (2024)

    2024

  24. [32]

    Kalvari, I. et al. Rfam 14: expanded coverage of metagenomic, viral and microRNA families. Nucleic Acids Res. 49, D192–D200 (2021)

    2021

  25. [33]

    Zou, S. et al. A Large-Scale Foundation Model for RNA Function and Structure Prediction. NeurIPS 2024 Workshop on AI for New Drug Modalities (2024)

    2024

  26. [34]

    T., Shahgir, H

    Tahmid, M. T., Shahgir, H. S., Mahbub, S., Dong, Y. & Bayzid, Md. S. BiRNA- BERT allows efficient RNA language modeling with adaptive tokenization. Preprint at https://doi.org/10.1101/2024.07.02.601703 (2024)

    2024 doi

  27. [35]

    Zhang, T. et al. RNAcmap: a fully automatic pipeline for predicting contact maps of RNAs by evolutionary coupling analysis. Bioinforma. Oxf. Engl. 37, 3494–3500 (2021)

    2021

  28. [36]

    & Liò, P

    Rossi, E., Monti, F., Bronstein, M. & Liò, P. ncRNA Classification with Graph Convolutional Networks. Preprint at http://arxiv.org/abs/1905.06515 (2019)

    2019 arXiv

  29. [37]

    & Yang, J

    Sun, S., Wang, W., Peng, Z. & Yang, J. RNA inter-nucleotide 3D closeness prediction by deep residual neural networks. Bioinforma. Oxf. Engl. 37, 1093–1098 (2021)

    2021

  30. [38]

    Xu, Y. et al. PrismNet: predicting protein-RNA interaction using in vivo RNA structural information. Nucleic Acids Res. 51, W468–W477 (2023)

    2023

  31. [39]

    Sample, P. J. et al. Human 5’ UTR design and variant effect prediction from a massively parallel translation assay. Nat. Biotechnol. 37, 803–809 (2019)

    2019

  32. [40]

    Cao, J. et al. High-throughput 5’ UTR engineering for enhanced protein production in non-viral gene therapies. Nat. Commun. 12, 4138 (2021)

    2021

  33. [41]

    & Gagneur, J

    Karollus, A., Avsec, Ž. & Gagneur, J. Predicting mean ribosome load for 5’UTR of any length using deep learning. PLoS Comput. Biol. 17, e1008982 (2021)

    2021

  34. [42]

    & Mathews, D

    Tan, Z., Fu, Y., Sharma, G. & Mathews, D. H. TurboFold II: RNA structural alignment and secondary structure prediction informed by multiple homologs. Nucleic Acids Res. 45, 11570–11581 (2017)

    2017

  35. [43]

    Sloma, M. F. & Mathews, D. H. Exact calculation of loop formation probability identifies folding motifs in RNA secondary structures. RNA N. Y. N 22, 1808–1818 (2016)

    2016

  36. [44]

    Danaee, P. et al. bpRNA: large-scale automated annotation and analysis of RNA secondary structure. Nucleic Acids Res. 46, 5381–5394 (2018)

    2018

  37. [45]

    Das, R. et al. Assessment of three-dimensional RNA structure prediction in CASP15. Proteins 91, 1747–1770 (2023)

    2023

  38. [46]

    & Zhou, Y

    Singh, J., Litfin, T., Singh, J., Paliwal, K. & Zhou, Y. SPOT-Contact-LM: improving single-sequence-based prediction of protein contact map using a transformer language model. Bioinformatics 38, 1888–1894 (2022)

    2022

  39. [47]

    Zhou, Z. et al. DNABERT-2: Efficient Foundation Model and Benchmark For Multi-Species Genome. Preprint at http://arxiv.org/abs/2306.15006 (2024)

    2024 arXiv

  40. [48]

    Dalla-Torre, H. et al. The Nucleotide Transformer: Building and Evaluating Robust Foundation Models for Human Genomics

  41. [49]

    & Whitehead, K

    Chaudhary, N., Weissman, D. & Whitehead, K. A. mRNA vaccines for infectious diseases: principles, delivery and clinical translation. Nat. Rev. Drug Discov. 20, 817– 838 (2021)

    2021

  42. [50]

    S., Goedel, A

    Rohner, E., Yang, R., Foo, K. S., Goedel, A. & Chien, K. R. Unlocking the promise of mRNA therapeutics. Nat. Biotechnol. 40, 1586–1600 (2022)

    2022

  43. [51]

    Garner, A. L. Contemporary Progress and Opportunities in RNA-Targeted Drug Discovery. ACS Med. Chem. Lett. 14, 251–259 (2023)

    2023

  44. [52]

    Childs-Disney, J. L. et al. Targeting RNA structures with small molecules. Nat. Rev. Drug Discov. 21, 736–762 (2022)

    2022

  45. [53]

    & Stadler, P

    Menzel, P., Gorodkin, J. & Stadler, P. F. The tedious task of finding homologous noncoding RNA genes. RNA 15, 2075–2082 (2009)

    2009

  46. [54]

    Sayers, E. W. et al. Database resources of the National Center for Biotechnology Information in 2023. Nucleic Acids Res. 51, D29–D38 (2023)

    2023

  47. [55]

    Chen, M. et al. Genome Warehouse: A Public Repository Housing Genome-Scale Data. Genomics Proteomics Bioinformatics 19, 584–589 (2021)

    2021

  48. [56]

    Martin, F. J. et al. Ensembl 2023. Nucleic Acids Res. 51, D933–D941 (2023)

    2023

  49. [57]

    Leebens-Mack, J. H. et al. One thousand plant transcriptomes and the phylogenomics of green plants. Nature 574, 679–685 (2019)

    2019

  50. [58]

    Haeussler, M. et al. The UCSC Genome Browser database: 2019 update. Nucleic Acids Res. 47, D853–D858 (2019)

    2019

  51. [59]

    & Fraternali, F

    Cavallo, L., Kleinjung, J. & Fraternali, F. POPS: A fast algorithm for solvent accessible surface areas at atomic and residue level. Nucleic Acids Res. 31, 3364–3366 (2003)

    2003

  52. [60]

    & Zhou, Y

    Singh, J., Hanson, J., Paliwal, K. & Zhou, Y. RNA secondary structure prediction using an ensemble of two-dimensional deep neural networks and transfer learning. Nat. Commun. 10, 5407 (2019)

    2019

  53. [61]

    Singh, J. et al. Improved RNA secondary structure and tertiary base-pairing prediction using evolutionary profile, mutational coupling and two-dimensional transfer learning. Bioinforma. Oxf. Engl. 37, 2589–2600 (2021)

    2021

  54. [62]

    Leontis, N. B. & Zirbel, C. L. Nonredundant 3D Structure Datasets for RNA Knowledge Extraction and Benchmarking. in RNA 3D Structure Analysis and Prediction (eds. Leontis, N. & Westhof, E.) 281–298 (Springer Berlin Heidelberg, Berlin, Heidelberg, 2012). doi:10.1007/978-3-642-2...

    2012 doi

  55. [63]

    & Zhou, Y

    Singh, J., Paliwal, K., Singh, J. & Zhou, Y. RNA Backbone Torsion and Pseudotorsion Angle Prediction Using Dilated Convolutional Neural Networks. J. Chem. Inf. Model. 61, 2610–2622 (2021)

    2021

  56. [64]

    Scalzitti, N. et al. Spliceator: multi-species splice site prediction using convolutional neural networks. BMC Bioinformatics 22, 561 (2021)

    2021

  57. [65]

    Song, Z. et al. Attention-based multi-label neural networks for integrated prediction and interpretation of twelve widely occurring RNA modifications. Nat. Commun. 12, 4011 (2021)

    2021

  58. [66]

    & Zhang, S.-W

    Han, Y. & Zhang, S.-W. ncRPI-LGAT: Prediction of ncRNA-protein interactions with line graph attention network framework. Comput. Struct. Biotechnol. J. 21, 2286– 2295 (2023)

    2023

  59. [67]

    Wen, M., Cong, P., Zhang, Z., Lu, H. & Li, T. DeepMirTar: a deep-learning approach for predicting human miRNA targets. Bioinforma. Oxf. Engl. 34, 3781–3787 (2018)

    2018

  60. [68]

    & Chen, M

    Kang, Q., Meng, J., Cui, J., Luan, Y. & Chen, M. PmliPred: a method based on hybrid model and fuzzy decision for plant miRNA–lncRNA interaction prediction. Bioinformatics 36, 2986–2992 (2020)

    2020

  61. [69]

    & Kulkarni, A

    Kolekar, P., Pataskar, A., Kulkarni-Kale, U., Pal, J. & Kulkarni, A. IRESPred: Web Server for Prediction of Cellular and Viral Internal Ribosome Entry Site (IRES). Sci. Rep. 6, 27436 (2016)

    2016

  62. [70]

    Weingarten-Gabbay, S. et al. Systematic discovery of cap-independent translation sequences in human and viral genomes. Science 351, aad4939 (2016)

    2016

  63. [71]

    Zhao, J. et al. IRESbase: a Comprehensive Database of Experimentally Validated Internal Ribosome Entry Sites. bioRxiv 2020.01.15.894592 (2020) doi:10.1101/2020.01.15.894592

    2020 doi

  64. [72]

    Seq.”: RNA sequence; “SS

    Mokrejš, M. et al. IRESite—a tool for the examination of viral and cellular internal ribosome entry sites. Nucleic Acids Res. 38, D131–D136 (2010). Author contributions Ya.Z., J.Z. and J.C. conceived and supervised the study. H.W. and Yi.Z. implemented the algorithms and perfo...

    2010

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.