The DOI matches its Crossref record; 93 of the 93 references it cites were found in a registry or an index; 9 of the 16 sources its claims cite were checked for support.
Evolutionary-scale prediction of atomic level protein structure with a language model
Open accessAbstract
Artificial intelligence has the potential to open insight into the structure of proteins at the scale of evolution. It has only recently been possible to extend protein structure prediction to two hundred million cataloged proteins. Characterizing the structures of the exponentially growing billions of protein sequences revealed by large scale gene sequencing experiments would necessitate a break-through in the speed of folding. Here we show that direct inference of structure from primary sequence using a large language model enables an order of magnitude speed-up in high resolution structure prediction. Leveraging the insight that language models learn evolutionary patterns across millions of sequences, we train models up to 15B parameters, the largest language model of proteins to date. As the language models are scaled they learn information that enables prediction of the three-dimensional structure of a protein at the resolution of individual atoms. This results in prediction that is up to 60x faster than state-of-the-art while maintaining resolution and accuracy. Building on this, we present the ESM Metage-nomic Atlas. This is the first large-scale structural characterization of metagenomic proteins, with more than 617 million structures. The atlas reveals more than 225 million high confidence predictions, including millions whose structures are novel in comparison with experimentally determined structures, giving an unprecedented view into the vast breadth and diversity of the structures of some of the least understood proteins on earth.
Overview
AI snapshot
Written from the paper’s full text.
TL;DR
The paper introduces ESM-2 protein language models up to 15B parameters and ESMFold, an end-to-end single-sequence structure predictor that reaches atomic-resolution accuracy while being up to 60x faster than state-of-the-art pipelines. It uses ESMFold to build the ESM Metagenomic Atlas, folding over 617 million MGnify90 sequences and identifying more than 225 million high-confidence structures, many novel relative to known sequences and structures.
Key findings
- Scaling ESM-2 from 8M to 15B parameters improves language modeling and structural information: validation perplexity falls from 10.45 to 6.37, unsupervised long-range contact precision (LR P@L) rises from 0.16 to 0.54, and structure-module-only LDDT rises from 0.48 to 0.72. The 15B model reaches TM-score 0.71 on CAMEO and 0.54 on CASP14, 0.064 higher than the 150M ESM-2 model on both.
- ESMFold achieves average TM-score 0.83 on CAMEO and 0.68 on CASP14, compared with AlphaFold2 at 0.88 and 0.85 and RosettaFold at 0.82 on CAMEO. On a single NVIDIA V100 GPU, ESMFold predicts a 384-residue protein in 14.2 seconds, 6x faster than a single AlphaFold2 model, with up to ~60x speedup on shorter sequences; MSA search alone can take over 10 minutes.
- The ESM Metagenomic Atlas folds over 617 million MGnify90 sequences in 2 weeks on ~2,000 GPUs, yielding ~365 million good-confidence predictions (mean pLDDT > 0.5 and pTM > 0.5) and ~225 million high-confidence predictions (mean pLDDT > 0.7 and pTM > 0.7). In a 1-million high-confidence sample, 76.8% have <90% sequence identity to UniRef90, 12.6% have no PDB structure with TM-score > 0.5, and 10.4% have both low structural similarity (TM-score ≤ 0.5) and no close sequence homolog (<30% identity).
- Confidence is well calibrated: for high-confidence predictions (pLDDT > 0.7), CAMEO accuracy is comparable to AlphaFold2 (ESMFold LDDT = 0.83, AlphaFold2 LDDT = 0.85); for very high confidence (pLDDT > 0.9), median all-atom RMSD95 is 1.42 Å and backbone RMSD95 is 0.94 Å.
Methods
- ESM-2 is a BERT-style transformer masked language model trained on UniRef50/UniRef90 sequences, scaled from 8M to 15B parameters, with models trained for up to 500K updates (15B for 270K updates) and evaluated on held-out UniRef50 clusters.
- ESMFold adds a 48-block folding head to a frozen 3B-parameter ESM-2, trained on ~25K PDB clusters covering ~325K experimentally determined chains plus ~12M AlphaFold2-predicted structures; evaluation uses temporally held-out CAMEO (194 proteins) and CASP14 (51 targets) test sets.
- Metagenomic analysis folded MGnify90 sequences of length 20–1024, covering 99% of the database, and used Foldseek, MMseqs2, blastp, and UMAP for structural and sequence similarity analyses.
Limitations
- The paper states that ESMFold was not trained on protein complexes, so complex prediction is out of distribution; it also notes that ablating MSAs for AlphaFold and RosettaFold is an artificial single-sequence setting because those models were not explicitly trained for single sequences.
- The paper states that calibration becomes critical as structure prediction scales, and that current models are far from the limit of scale in parameters, sequence data, and compute.
- Evident from the text: this is a bioRxiv preprint not certified by peer review; ESMFold is less accurate than AlphaFold2 overall on CAMEO (0.83 vs 0.88 TM-score) and CASP14 (0.68 vs 0.85), with a larger gap on hard CAMEO targets (0.45 vs 0.62).
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
9 of 16 cited sources checked · 5 supported · 0 flagged · accuracy not yet measured
Every source the paper cites for this claim was checked and supports it.
In the paper
“the vast majority (76.8%) of high confidence predictions being separate from UniRef90 (32) by at least 90% sequence identity”
What the sources say
- Supportedcites [32] source
The citation is attached to UniRef90, identifying the reference database used for the citing paper’s comparison. The Introduction describes UniRef90 as clustering sequences at the 90% identity level; the 76.8% figure is the citing paper’s own result, not a result attributed to this source. (LLM, conf=0.90)
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Very high confidence predictions in the metagenomic atlas are expected to often be reliable at a resolution sufficient for insight similar to experimentally determined structures, such as into the biochemistry of active sites (61);”
What the sources say
- Partly supportedcites [61] sourceExperimental — accuracy not yet measured
The High-resolution models section says predicted models can meet demanding structural requirements and sometimes support ligand-binding studies; the Medium-resolution models section says predictions can help locate active sites. The paper does not assess predictions in the metagenomic atlas or establish that its very high-confidence predictions are often reliable enough to provide experimentally comparable insight into active-site biochemistry. Page numbers are unavailable in our copy. (LLM, conf=0.85)
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“We fold over 617 million sequences from the MGnify90 database (25).”
What the sources say
- Partly supportedcites [25] sourceExperimental — accuracy not yet measured
The 'UPDATES TO THE MGnify PROTEIN DATABASE' section describes MGnify protein sequences clustered at 90% identity, supporting the identification of the database. It reports over 1.1 billion sequences in the full database, but does not give a count of 617 million for the 90%-clustered set. The citing authors' folding of those sequences is their own work, not a finding attributed to this source. (LLM, conf=0.84)
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“In a comparison with AlphaFold-Multimer (45) on a dataset of 2,978 recent multimeric complexes deposited in the PDB, ESMFold achieves the same qualitative DockQ (39) categorization for 53.2% of chain pairs, despite not being trained on protein complexes (Fig. S6).”
What the sources say
- Supportedcites [45] Richard Evans , Michael O’Neill , Alexander Pritzel , Natasha Antropova , Andrew Senior , Tim Green , Augustin Žídek , Russ Bates , Sam Blackwell , Jason Yim , Olaf Ronneberger , Sebastian Bodenstein , Michal Zielinski , Alex Bridgland , Anna Potapenko , Andrew Cowie , Kathryn Tunyasuvunakool , Rishub Jain , Ellen Clancy , Pushmeet Kohli , John Jumper , and Demis Hassabis . Protein complex prediction with AlphaFold-Multimer, March 2022. URL https://www.biorxiv.org/content/10.1101/2021.10.04.463034v2. Pages: 2021.10.04.463034 Section: New Results. source
The citation identifies AlphaFold-Multimer as the comparator. The Results (p. 5) describe this paper's AlphaFold-Multimer predictions and benchmarks. The 2,978-complex comparison and 53.2% figure are presented as the citing authors' own results, not findings attributed to this source. (LLM, conf=0.92)
- Supportedcites [39] source
Citation (39) is attached to “DockQ” and identifies the quality measure used for the categorization. The Conclusions describe DockQ as a protein–protein docking model quality score that closely recapitulates the four CAPRI quality classes; the citation is not offered as the source of the ESMFold comparison or its 53.2% result. (LLM, conf=0.93)
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- Supported
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This training also results in the emergence of structure in the models.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesA result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Here we show that direct inference of structure from primary sequence using a large language model enables an order of magnitude speed-up in high resolution structure prediction.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesA result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We discover that an atomic resolution structure prediction can be projected from the representations of the ESM-2 language models.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesThe paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“We are able to complete this characterization in 2 weeks on a heterogeneous cluster of 2,000 GPUs”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesThe paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This results in an improvement in speed of up to 60x on the inference forward pass alone”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesA result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The atlas reveals more than 225 million high confidence predictions”
- Also stated in: § Evolutionary-scale prediction of atomic level protein structure with a language model · p. 2 · plain
“High confidence predictions are made for over 225M structures”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- Also stated in: § Evolutionary-scale prediction of atomic level protein structure with a language model · p. 2 · plain
- It may state two separate claims as one. 32
- Probably not a substantial claim. 27
- The quote only partly states this claim. 1
- Borderline: one of the tests for a substantial claim is unclear. 1
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“This is the first large-scale structural characterization of metagenomic proteins, with more than 617 million structures.”
- Also stated in: § Evolutionary-scale prediction of atomic level protein structure with a language model · p. 2 · plain
“We present the first evolutionary scale structural characterization of a metagenomic resource, folding practically all sequences in MGnify90 (25), over 617M proteins.”
What the sources say
- Could not verifycites [25] sourceExperimental — accuracy not yet measured
The ‘UPDATES TO THE MGnify PROTEIN DATABASE’ section describes more than 1.1 billion protein sequences and their clustering, not predicted structures. Searches for the ESM Metagenomic Atlas, 617 million structures, and protein folding found no passage addressing the claimed structural characterization. (LLM, conf=0.76)
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
- Also stated in: § Evolutionary-scale prediction of atomic level protein structure with a language model · p. 2 · plain
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Our second test set consists of 51 targets from the CASP14 competition (37).”
What the sources say
- Partly supportedcites [37] sourceExperimental — accuracy not yet measured
The Introduction describes CASP14 modeling targets, supporting their use as a source for a test set. It reports 68 tertiary-structure modeling targets, but does not identify the citing authors’ 51-target selection or verify its size. (LLM, conf=0.85)
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“MGnify sequence MGYP000936678158 has no significant sequence matches to any entry in UniRef90, nor any significant matches via a jackhmmer (47) reference proteome search, but has a predicted structure conserved across many nucleases (PDB 5YET B, TM-score 0.68; PDB 3HR4 A, TM-score 0.67)”
What the sources say
- Supportedcites [47] source
Split into 2 atoms — 2 confirmed, 0 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“During method development (e.g. hyperparameter selection), we used a temporally held out validation set obtained from the Continuous Automated Model EvaluatiOn (CAMEO) server (36) by filtering from August 2021 to January 2022.”
What the sources say
- Supportedcites [36] source
The citation identifies the CAMEO platform from which the citing authors obtained data. The Introduction describes CAMEO’s automated evaluations and says its benchmarking results are publicly available for methods development. The August 2021–January 2022 filtering is the citing authors’ own procedure, not a claim attributed to this 2018 source. (LLM, conf=0.92)
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
- Supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The resulting ESM-2 model family significantly outperforms previously state-of-the-art ESM-1b (a ∼ 650 million parameter model) at a comparable number of parameters, and on structure prediction benchmarks it also outperforms other recent protein language models (Table S3).”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“At 270K training steps the 8M parameter model has a perplexity of 10.45, and the 15B model reaches a perplexity of 6.37”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This results in prediction that is up to 60x faster than state-ofthe-art while maintaining resolution and accuracy.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This removes costly aspects of current state-of-the-art structure prediction pipelines, eliminating the need for a multiple sequence alignment, while at the same time greatly simplifying the neural architecture used for inference.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Attention patterns develop in ESM-2 that correspond to tertiary structure (Fig. 1A), and scaling leads to large improvements in the understanding of structure (Fig. 1B).”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“and ∼ 225 million predictions with high confidence (mean pLDDT > 0 . 7 and pTM > 0 . 7 ) corresponding to ∼ 36% of total structures folded (Fig. 3).”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The first test is from CAMEO, consisting of all 194 test proteins from April 01, 2022 through June 25, 2022.”
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This results in a speed improvement for structure prediction of more than an order of magnitude while maintaining high resolution accuracy.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We were able to complete the predictions in 2 weeks on a cluster of approximately 2,000 GPUs (Appendix A.4.1).”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“On the CAMEO test set, language model perplexity has a Pearson correlation of -0.55 with the TM-score between the predicted and experimental structures; on CASP14, the correlation is -0.67 (Fig. 2B).”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This results in 25,450 clusters covering a total of 325,498 chains.”
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“ESMFold achieves competitive accuracy with RosettaFold on CAMEO, which averages a TM-score of 0.82.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The average long range P@L was 0.4287 with a standard deviation of 0.0028.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesThe quote only partly states this claim.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“MGnify sequence MGYP004000959047 has no significant UniRef90 or jackhmmer reference proteome matches but its predicted structure has high similarity to experimental structures of lipid binding domains (PDB 6BYM A, TMscore 0.80; PDB 5YQP B, TM-score 0.78)”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Using the search protocols released with AlphaFold2, including MSAs and templates, AlphaFold2 achieves 0.88 and 0.85 on CAMEO and CASP14 respectively.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“For Hsu et al. (62) we generated a set of 13,477,259 structure predictions with AlphaFold2 using MSAs generated via the process in Rao et al. (40).”
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The 15 billion parameter model reaches a TM-score of 0.71 on the CAMEO test set and 0.54 on the CASP14 test set, 0.064 points higher than the 150 million parameter ESM-2 model on both (Fig. 1E).”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This characterization reveals the structures of hundreds of millions proteins that have been previously unknown, millions of which are novel in comparison to experimentally determined structures.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“ESMFold achieves an average TM-score of 0.83 on CAMEO and 0.68 on CASP14.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“No filtering is performed on these test sets, as ESMFold is able to make predictions on all sequences, including the length-2166 target T1044.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Because of the way the dataset is constructed, only 1.5% of the dataset is removed with this filter.”
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Language models greatly accelerate state-of-the-art high resolution structure prediction.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes? Based: unclearBorderline substantial claim. It is unclear why the paper says it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“On a single NVIDIA V100 GPU, ESMFold makes a prediction on a protein with 384 residues in 14.2 seconds, 6x faster than a single AlphaFold2 model.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Across all models (Table S3) there is a correlation of -0.99 between validation perplexity and CASP14 TM-score, and -1.00 between validation perplexity and CAMEO TM-score”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“On the CAMEO test set, ESMFold predictions have a median all-atom RMSD 95 (root-mean-squared deviation at 95% residue coverage) of 1.91 ˚ A and backbone RMSD 95 of 1.33 ˚ A.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“For RoseTTAfold, the speed of the SE(3) Transformer dominates, especially at low sequence lengths.”
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“These steps result in a total of 2978 clusters.”
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“At low sequence lengths, ESMFold is dominated by language model performance, while the O ( N 3 ) computation of pairwise representations takes over at high sequence lengths.”
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Differentiation is greater in the ªmediumº and ªhardº classes, which have lower MSA depth and are better predicted by AlphaFold2.”
Strength:plain· not yet calibratedCausal:correlational· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We see an over 60x speed advantage for shorter protein sequences, and a reasonable speed advantage for longer protein sequences.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“which we were able to do in 2 weeks time.”
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Unsupervised contact prediction results show that both RoPE and newer data significantly improve the results.”
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“We estimate that on a homogeneous network GPU cluster of V100s, the entire 620 million sequences would take approximately 28,000 GPU days to fold”
Strength:hedged· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Language model perplexity and ESMFold pLDDT are both well correlated with actual structure prediction accuracy on CASP14 and CAMEO.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“After 270K updates the 8M parameter model has a perplexity of 10.45, and the 15B model reaches a perplexity of 6.37.”
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“After filtering, there are 104 sequences with MSA depth ≤ 100 , 70 sequences with MSA depth ≤ 10 , and 22 sequences with MSA depth = 1.”
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“analyzed the high-confidence subset with mean pLDDT greater than 0.9, corresponding to ∼ 59K structures.”
Strength:plain· not yet calibratedCausal:correlational· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This results in a total of 3505 unique chain pairs.”
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We do observe a slight regression when sampling from UniRef90 clusters”
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“These capabilities develop as a function of scale, with greater capabilities emerging as computation, data, and number of parameters increase.”
What the sources say
- Not matchedcites 19±22
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
- Not matched
A claim is checked only against the cited paper’s full text, never its abstract, and none is available for its sources.
In the paper
“Within a protein family, structure and function can be inferred from the patterns in sequences (4, 5).”
What the sources say
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This means that improvement on sequences with high evolutionary depth saturates at lower model scales, and improvement on sequences with low evolutionary depth continues as models increase in size.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
In the paper
“language models of text, despite the simplicity of their training objectives, such as filling in missing words or predicting the next word, are shown to develop emergent capabilities that are connected to the underlying meaning of the text.”
What the sources say
- Not matchedcites 19±22
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“As the language models are scaled they learn information that enables prediction of the three-dimensional structure of a protein at the resolution of individual atoms.”
- Also stated in: § Evolutionary-scale prediction of atomic level protein structure with a language model · p. 2 · plain
“We discover that atomic resolution structure prediction emerges and continues to improve in language models over the four orders of magnitude in parameter scale.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
- Also stated in: § Evolutionary-scale prediction of atomic level protein structure with a language model · p. 2 · plain
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“No filtering is performed on these test sets, as ESMFold is able to make predictions on all sequences, including the length-2166 target T1044.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“All train sequences which match a validation sequence with 50% sequence identity under this search are removed from the train set.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“We report results by testing 3D structure prediction models on two test sets, both chosen to be temporally held out from our supervised training set.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“All compared sequences are temporally held out from the training set.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“however we believe this difference is small and the UniRef90 cluster sampling is likely to help for the larger models.”
Strength:hedged· not yet calibratedCausal:conditional· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✗ Based: failsProbably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“The number of SE(3) max-iterations are artificially limited to 20 (default 200) and no MSAs are used as input for these measurements.”
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“Additionally, this only measures the network forward time, and does not include the time taken to compute sidechains with PyRosetta or search for MSAs.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim. It may state two separate claims as one.
In the paper
“Since BERT style models have been shown to be amenable to very large batch sizes (67), we increased our effective batch size to 2M tokens.”
What the sources say
- Not matchedcites [67] Yang You , Jing Li , Sashank Reddi , Jonathan Hseu , Sanjiv Kumar , Srinadh Bhojanapalli , Xiaodan Song , James Demmel , Kurt Keutzer , and Cho-Jui Hsieh. Large Batch Optimization for Deep Learning: Training BERT in 76 Minutes. page 38, 2020.
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Strength:boosted· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
- Not matched
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“even though ESMFold is not trained on protein complexes.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
The paper relies on this without support.
AlphaFold2 predictions provide a valid reference for assessing the accuracy of ESMFold predictions on metagenomic proteins.
“We also assess calibration against AlphaFold predictions on metagenomic proteins.”
“On a random subset of ∼ 4K metagenomic sequences, there is a high correlation (Pearson r = 0.79) between ESMFold pLDDT and the LDDT to AlphaFold2 predictions (Fig. 3A).”
The paper relies on this without support.
The MGnify90 database is representative of metagenomic protein diversity.
“We fold over 617 million sequences from the MGnify90 database (25).”
The paper relies on this without support.
The PDB database provides a comprehensive reference for experimentally determined protein structures.
“no structure is found with TM-score over 0.5 in the PDB database”
“no experimentally determined structure with a similar fold could be identified”
The paper relies on this without support.
Absolute positional encoding methods don't extrapolate well beyond the context window they are trained on.
“However, absolute positional encoding methods don't extrapolate well beyond the context window they are trained on.”
The paper relies on this without support.
Orphan proteins are very challenging for current structure prediction models due to a lack of evolutionary information.
“Due to a lack of evolutionary information, these sequences can be very challenging for current structure prediction models.”
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“Evo also reaches competitive performance with leading protein-specific language models (41, 46–48) (Fig. 2B).”
From Sequence modeling and design from molecular to genome scale with Evo (2024) · cites this paper as [47]
The cited source says what the claim attributes to it.
In the citing paper
“Such modeling paradigms have been used successfully in the biomolecular realm, for example, to predict three-dimensional (3D) molecular structures from sequences32–34 and interactions between different molecular components.35–38”
From How to build the virtual cell with artificial intelligence: Priorities and opportunities (2024) · cites this paper as [34]
In the citing paper
“are the best and most accurate answer to the structure prediction problem to date”
From The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins (2024) · cites this paper as [19]
In the citing paper
“The unprecedented accuracy of AlphaFold2,127 RosettaFold,128 and others129,130 offers, at least in principle, the opportunity to virtually screen any protein of interest with the structure-based approach, irrespective of its a priori ligandability (after AlphaFold2, the structural coverage of the human proteome increased from 48% to 76% of the total residues).131”
From Advancing Targeted Protein Degradation via Multiomics Profiling and Artificial Intelligence (2023) · cites this paper as [129]
In the citing paper
“This difficulty is amplified by the fact that the repository of publicly available structures, consisting of the combined databases of AFDB and the ESMatlas38, is approaching a billion entries.”
From Clustering predicted structures at the scale of the known protein universe (2023) · cites this paper as [38]
In the citing paper
“the ESM Atlas contains over 617 million metagenomic structures predicted by ESMFold4.”
From Fast and accurate protein structure search with Foldseek (2024) · cites this paper as [4]
In the citing paper
“Our current proof-of-concept model is considerably smaller—by 200 times—than the largest published protein language model (49).”
From DNA language models are powerful predictors of genome-wide variant effects (2023) · cites this paper as [49]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Conflicting PDB identifier for the nuclease example
MinorThe main text identifies the comparison structure as PDB 3HR4 A, but the Figure 4 caption identifies it as PDB 3H4R. These are different identifiers for the same described example.
“PDB 3HR4 A, TM-score 0.67”
“The predicted structure of MGYP000936678158 aligns to an experimental structure from a bacterial nuclease (light brown, PDB: 3H4R)”
Citations
Cites
Basic local alignment search tool
Altschul, S. F., Gish, W., Miller, W., Myers, E. W., & Lipman, D. J. (1990) · Journal of Molecular Biology · doi:10.1006/jmbi.1990.9999 · Cited by 98,669
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018) · arXiv · doi:10.48550/arxiv.1810.04805 · Cited by 73,173
Scikit-learn: Machine Learning in Python
Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Müller, A., Nothman, J., Louppe, G., et al. (2012) · arXiv · doi:10.48550/arxiv.1201.0490 · Cited by 71,016
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 54,325
Highly accurate protein structure prediction with AlphaFold
Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Žídek, A., Potapenko, A., et al. (2021) · Nature · doi:10.1038/s41586-021-03819-2 · Cited by 43,764
Language Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372
A Programmable Dual-RNA–Guided DNA Endonuclease in Adaptive Bacterial Immunity
Jinek, M., Chylinski, K., Fonfara, I., Hauer, M., Doudna, J. A., & Charpentier, E. (2012) · Science · doi:10.1126/science.1225829 · Cited by 17,116
ColabFold - Making protein folding accessible to all
Mirdita, M., Schütze, K., Moriwaki, Y., Heo, L., Ovchinnikov, S., & Steinegger, M. (2021) · doi:10.1101/2021.08.15.456425 · Cited by 9,391
SCANPY: large-scale single-cell gene expression data analysis
Wolf, F. A., Angerer, P., & Theis, F. J. (2018) · Genome Biology · doi:10.1186/s13059-017-1382-0 · Cited by 8,881
Varadi, M., Anyango, S., Deshpande, M., Nair, S., Natassia, C., Yordanova, G., Yuan, D., Stroe, O., Wood, G., Laydon, A., et al. (2022) · Nucleic Acids Research · doi:10.1093/nar/gkab1061 · Cited by 7,850
Flagged first, then by citations
Research works printed but not linked: 1
- Ashish Vaswani , Noam Shazeer , Niki Parmar , Jakob Uszkoreit , Llion Jones , Aidan N Gomez , Łukasz Kaiser , and Illia Polosukhin . Attention Is All You Need. In Advances in Neural Information Processing Systems, pages 5998–6008, 2017. URL https://papers.nips.cc/paper/7181-attention-is-all-you-need.pdf.
Cited by
Sorted by citations
Items 101–200 of 5,664
GenSLMs: Genome-scale language models reveal SARS-CoV-2 evolutionary dynamics
Zvyagin, M., Brace, A., Hippe, K., Deng, Y., Zhang, B., Bohorquez, C. O., Clyde, A., Kale, B., Perez-Rivera, D., Ma, H., et al. (2022) · doi:10.1101/2022.10.10.511571 · Cited by 120
A foundation model of transcription across human cell types
(2025) · doi:10.1038/s41586-024-08391-z · Cited by 130
The social and structural architecture of the yeast protein interactome
Michaelis, A. C., Brunner, A.-D., Zwiebel, M., Meier, F., Strauss, M. T., Bludau, I., & Mann, M. (2023) · Nature · doi:10.1038/s41586-023-06739-5 · Cited by 119
“We envisioned that recent breakthroughs in protein structure prediction 44–46 , including the ability to model several subunits of protein complexes 43,47 , would add an additional dimension to our interactome dataset.”
BackgroundDesign of highly functional genome editors by modelling CRISPR–Cas sequences
Ruffolo, J. A., Nayfach, S., Gallagher, J. P., Bhatnagar, A., Beazer, J., Hussain, R., Russ, J., Yip, J., Hill, E., Pačesa, M., et al. (2025) · Nature · doi:10.1038/s41586-025-09298-z · Cited by 233
Atom-level enzyme active site scaffolding using RFdiffusion2
Ahern, W., Yim, J., Tischer, D., Salike, S., Woodbury, S. M., Kim, D., Kalvet, I., Kipnis, Y., Coventry, B., Altae-Tran, H., et al. (2025) · Nature Methods · doi:10.1038/s41592-025-02975-x · Cited by 157
Genomic language models: opportunities and challenges
Benegas, G., Ye, C., Albors, C., Li, J. C., & Song, Y. S. (2025) · Trends in Genetics · doi:10.1016/j.tig.2024.11.013 · Cited by 157
A generalized platform for artificial intelligence-powered autonomous enzyme engineering
Singh, N., Lane, S., Yu, T., Lu, J., Ramos, A., Cui, H., & Zhao, H. (2025) · Nature Communications · doi:10.1038/s41467-025-61209-y · Cited by 124
Sparks of function by de novo protein design
Chu, A. E., Lu, T., & Huang, P. (2024) · Nature Biotechnology · doi:10.1038/s41587-024-02133-2 · Cited by 131
Current and future directions in network biology
Žitnik, M., Li, M. M., Wells, A., Glass, K., Gysi, D. M., Krishnan, A., Murali, T. M., Predrag Radivojac, A., Roy, S., Baudot, A., et al. (2024) · Bioinformatics Advances · doi:10.1093/bioadv/vbae099 · Cited by 128
Disease diagnostics using machine learning of B cell and T cell receptor sequences
Zaslavsky, M., Craig, E., Michuda, J., Sehgal, N., Ram-Mohan, N., Lee, J.-Y., Nguyen, K. D., Hoh, R. A., Pham, T. D., Röltgen, K., et al. (2025) · Science · doi:10.1126/science.adp2407 · Cited by 132
A survey of generative AI for de novo drug design: new frontiers in molecule and protein generation
Tang, X., Dai, H., Knight, E., Wu, F., Li, Y., Li, T., & Gerstein, M. (2024) · Briefings in Bioinformatics · doi:10.1093/bib/bbae338 · Cited by 150
Robust enzyme discovery and engineering with deep learning using CataPro
Wang, Z., Xie, D., Wu, D., Luo, X., Wang, S., Li, Y., Yang, Y., Li, W., & Zheng, L. (2025) · Nature Communications · doi:10.1038/s41467-025-58038-4 · Cited by 123
The state of quantum computing applications in health and medicine
Flöther, F. F. (2023) · Research Directions Quantum Technologies · doi:10.1017/qut.2023.4 · Cited by 128
Foundation models in bioinformatics
Guo, F., Guan, R., Li, Y., Liu, Q., Wang, X., Yang, C., & Wang, J. (2025) · National Science Review · doi:10.1093/nsr/nwaf028 · Cited by 127
A new age in protein design empowered by deep learning
Khakzad, H., Igashov, I., Schneuing, A., Goverde, C. A., Bronstein, M. M., & Correia, B. E. (2023) · Cell Systems · doi:10.1016/j.cels.2023.10.006 · Cited by 121
Wang, J., Feng, J., Kang, Y., Pan, P., Ge, J., Wang, Y., Wang, M., Wu, Z., Zhang, X., Yu, J., et al. (2025) · Science Advances · doi:10.1126/sciadv.ads8932 · Cited by 125
Generalized Biomolecular Modeling and Design with RoseTTAFold All-Atom
Krishna, R., Wang, J., Ahern, W., Sturmfels, P., Venkatesh, P., Kalvet, I., Lee, G. R., Morey-Burrows, F. S., Anishchenko, I., Humphreys, I. R., et al. (2023) · doi:10.1101/2023.10.09.561603 · Cited by 106
Markus, B., Gruber Christian, C., Andreas, K., Kummer, A., Stefan, L., Oberdorfer, G., Elina, S., & Šnajdrová, R. (2023) · ACS Catalysis · doi:10.1021/acscatal.3c03417 · Cited by 121
CatPred: a comprehensive framework for deep learning in vitro enzyme kinetic parameters
Boorla, V. S., & Maranas, C. D. (2025) · Nature Communications · doi:10.1038/s41467-025-57215-9 · Cited by 122
Structure prediction of alternative protein conformations
Bryant, P., & Noé, F. (2024) · Nature Communications · doi:10.1038/s41467-024-51507-2 · Cited by 105
xTrimoPGLM: Unified 100B-Scale Pre-trained Transformer for Deciphering the Language of ProteinPreprint
Chen, B., Cheng, X., Li, P., Geng, Y.-A., Gong, J., Li, S., Bei, Z.-L., Tan, X., Wang, B., Zeng, X., et al. (2024) · bioRxiv · doi:10.1101/2023.07.05.547496 · Cited by 157
McCafferty, C. L., Klumpe, S., Amaro, R. E., Kukulski, W., Collinson, L., & Engel, B. D. (2024) · Cell · doi:10.1016/j.cell.2024.01.005 · Cited by 113
ChatGPT, GPT-4, and Other Large Language Models: The Next Revolution for Clinical Microbiology?
Egli, A. (2023) · Clinical Infectious Diseases · doi:10.1093/cid/ciad407 · Cited by 109
Data splitting to avoid information leakage with DataSAIL
Joeres, R., Blumenthal, D. B., & Kalinina, O. V. (2025) · Nature Communications · doi:10.1038/s41467-025-58606-8 · Cited by 159
Functional and evolutionary significance of unknown genes from uncultivated taxa
del Río, Á. R., Giner‐Lamia, J., Cantalapiedra, C. P., Botas, J., Deng, Z., Hernández-Plaza, A., Palmer, M. M., Santamaría‐Hernando, S., Rodríguez‐Herva, J. J., Ruscheweyh, H., et al. (2023) · Nature · doi:10.1038/s41586-023-06955-z · Cited by 110
Moi, D., Bernard, C., Steinegger, M., Nevers, Y., Langleib, M., & Dessimoz, C. (2025) · Nature Structural & Molecular Biology · doi:10.1038/s41594-025-01649-8 · Cited by 110
From task structures to world models: what do LLMs know?
Yildirim, I., & Paul, L. A. (2024) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2024.02.008 · Cited by 132
Fang, Y., Jiang, Y. F., Wei, L., Ma, Q., Ren, Z., Yuan, Q., & Wei, D. (2023) · Bioinformatics · doi:10.1093/bioinformatics/btad718 · Cited by 103
A Transformer-Based Ensemble Framework for the Prediction of Protein–Protein Interaction Sites
Mou, M., Pan, Z., Zhou, Z., Zheng, L., Zhang, H., Shi, S., Li, F., Sun, X., & Zhu, F. (2023) · Research · doi:10.34133/research.0240 · Cited by 106
Jiang, P., Sonne, C., Li, W., You, F., & You, S. (2024) · Engineering · doi:10.1016/j.eng.2024.04.002 · Cited by 130
Tesson, F., Planel, R., Egorov, A. A., Georjon, H., Vaysset, H., Brancotte, B., Néron, B., Mordret, E., Atkinson, G. C., Bernheim, A., et al. (2024) · Peer Community Journal · doi:10.24072/pcjournal.470 · Cited by 120
Cao, D., Chen, M., Zhang, R., Wang, Z., Huang, M., Yu, J., Jiang, X., Fan, Z., Zhang, W., Zhou, H., et al. (2025) · Nature Methods · doi:10.1038/s41592-024-02516-y · Cited by 106
Unrealized targets in the discovery of antibiotics for Gram-negative bacterial infections
Theuretzbacher, U., Blasco, B., Duffey, M., & Piddock, L. J. V. (2023) · Nature Reviews Drug Discovery · doi:10.1038/s41573-023-00791-6 · Cited by 105
Geometric deep learning for drug discovery
Liu, M., Li, C., Chen, R., Cao, D., & Zeng, X. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.122498 · Cited by 98
PLMSearch: Protein language model powers accurate and fast sequence search for remote homology
Liu, W., Wang, Z., You, R., Xie, C., Wei, H., Xiong, Y., Yang, J., & Zhu, S. (2024) · Nature Communications · doi:10.1038/s41467-024-46808-5 · Cited by 96
AlphaFold, allosteric, and orthosteric drug discovery: Ways forward
Nussinov, R., Zhang, M., Liu, Y., & Jang, H. (2023) · Drug Discovery Today · doi:10.1016/j.drudis.2023.103551 · Cited by 96
Diverse anti-defence systems are encoded in the leading region of plasmids
Samuel, B., Mittelman, K., Croitoru, S. Y., Haim, M. B., & Burstein, D. (2024) · Nature · doi:10.1038/s41586-024-07994-w · Cited by 97
Genomic language model predicts protein co-regulation and function
Hwang, Y., Cornman, A. L., Kellogg, E. H., Ovchinnikov, S., & Girguis, P. R. (2024) · Nature Communications · doi:10.1038/s41467-024-46947-9 · Cited by 99
Protein codes promote selective subcellular compartmentalization
Kilgore, H. R., Chinn, I., Mikhael, P. G., Mitnikov, I., Van Dongen, C., Zylberberg, G., Afeyan, L., Banani, S. F., Wilson-Hawken, S., Lee, T. I., et al. (2025) · Science · doi:10.1126/science.adq2634 · Cited by 99
Building Enzymes through Design and Evolution
Hossack, E. J., Hardy, F. J., & Green, A. P. (2023) · ACS Catalysis · doi:10.1021/acscatal.3c02746 · Cited by 102
Integrating protein language models and automatic biofoundry for enhanced protein evolution
Zhang, Q., Chen, W., Qin, M., Wang, Y., Pu, Z., Ding, K., Liu, Y., Zhang, Q., Li, D., Li, X., et al. (2025) · Nature Communications · doi:10.1038/s41467-025-56751-8 · Cited by 101
Yuan, Q., Xie, J., Xie, J., Zhao, H., & Yang, Y. (2023) · Briefings in Bioinformatics · doi:10.1093/bib/bbad117 · Cited by 93
pLM4ACE: A protein language model based predictor for antihypertensive peptide screening
Du, Z., Ding, X., Hsu, W. W. Y., Munir, A., Xu, Y., & Li, Y. (2023) · Food Chemistry · doi:10.1016/j.foodchem.2023.137162 · Cited by 98
An all-atom protein generative model
Chu, A. E., Kim, J., Cheng, L., Nesr, G. E., Xu, M., Shuai, R. W., & Huang, P. (2024) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2311500121 · Cited by 104
Advancing Ligand Docking through Deep Learning: Challenges and Prospects in Virtual Screening
Zhang, X., Shen, C., Zhang, H., Kang, Y., Hsieh, C., & Hou, T. (2024) · Accounts of Chemical Research · doi:10.1021/acs.accounts.4c00093 · Cited by 94
Graph Neural Networks in Modern AI-Aided Drug Discovery
Zhang, O., Lin, H., Zhang, X., Wang, X., Wu, Z., Ye, Q., Zhao, W., Wang, J., Ying, K., Kang, Y., et al. (2025) · Chemical Reviews · doi:10.1021/acs.chemrev.5c00461 · Cited by 110
Design of highly functional genome editors by modeling the universe of CRISPR-Cas sequences
Ruffolo, J. A., Nayfach, S., Gallagher, J., Bhatnagar, A., Beazer, J., Hussain, R., Russ, J., Yip, J., Hill, E., Pacesa, M., et al. (2024) · doi:10.1101/2024.04.22.590591 · Cited by 90
A Comprehensive Survey of Foundation Models in Medicine
Khan, W., Leem, S., See, K. B., Wong, J. K., Zhang, S., & Fang, R. (2025) · IEEE Reviews in Biomedical Engineering · doi:10.1109/rbme.2025.3531360 · Cited by 137
Sala, D., Hildebrand, P. W., & Meiler, J. (2023) · Frontiers in Molecular Biosciences · doi:10.3389/fmolb.2023.1121962 · Cited by 80
Democratizing protein language models with parameter-efficient fine-tuning
Sledzieski, S., Kshirsagar, M., Baek, M., Dodhia, R., Lavista Ferres, J., & Berger, B. (2024) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2405840121 · Cited by 86
Bittrich, S., Bhikadiya, C., Bi, C., Chao, H., Duarte, J. M., Dutta, S., Fayazi, M., Henry, J., Khokhriakov, I., Lowe, R., et al. (2023) · Journal of Molecular Biology · doi:10.1016/j.jmb.2023.167994 · Cited by 85
Qiu, X., Li, H., Steeg, G. V., & Godzik, A. (2024) · Biomolecules · doi:10.3390/biom14030339 · Cited by 103
Yao, L., Guan, J., Xie, P., Chung, C., Zhao, Z., Dong, D., Guo, Y., Zhang, W., Deng, J., Pang, Y., et al. (2024) · Nucleic Acids Research · doi:10.1093/nar/gkae1019 · Cited by 101
The Large Language Models on Biomedical Data Analysis: A Survey
Lan, W., Tang, Z., Liu, M., Chen, Q., Peng, W., Chen, Y. P., & Pan, Y. F. (2025) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2025.3530794 · Cited by 100
The future of chemistry is language
White, A. D. (2023) · Nature Reviews Chemistry · doi:10.1038/s41570-023-00502-0 · Cited by 88
Qiu, Y., & Wei, G. (2023) · Briefings in Bioinformatics · doi:10.1093/bib/bbad289 · Cited by 90
Zhang, X., Ibrahim, Z., Khaskheli, M. B., Raza, H., Zhou, F., & Shamsi, I. H. (2024) · Sustainability · doi:10.3390/su16177651 · Cited by 93
Koh, H. Y., Zheng, Y., Yang, M., Arora, R., Webb, G. I., Pan, S., Li, L., & Church, G. M. (2025) · Nature Reviews Bioengineering · doi:10.1038/s44222-025-00349-8 · Cited by 117
Predicting cellular responses to perturbation across diverse contexts with State
Adduri, A. K., Gautam, D., Bevilacqua, B., Naghipourfar, M., Imran, A., Shah, R., Teyssier, N., Verma, R., Carpenter, C., Eraslan, B., et al. (2026) · Cell · doi:10.1016/j.cell.2026.07.052 · Cited by 100
Generative artificial intelligence for de novo protein design
Winnifrith, A., Outeiral, C., & Hie, B. (2024) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2024.102794 · Cited by 103
Self-citationBFVD—a large repository of predicted viral protein structures
Kim, R. S., Levy Karin, E., Mirdita, M., Chikhi, R., & Steinegger, M. (2025) · Nucleic Acids Research · doi:10.1093/nar/gkae1119 · Cited by 86
Hierarchical graph transformer with contrastive learning for protein function prediction
Gu, Z., Luo, X., Chen, J., Deng, M., & Lai, L. (2023) · Bioinformatics · doi:10.1093/bioinformatics/btad410 · Cited by 92
Wang, R., Wang, T., Zhuo, L., Wei, J., Fu, X., Zou, Q., & Yao, X. (2024) · Briefings in Bioinformatics · doi:10.1093/bib/bbae078 · Cited by 93
Ding, K., Chin, M., Zhao, Y., Huang, W., Mai, B. K., Wang, H., Liu, P., Yang, Y., & Luo, Y. (2024) · Nature Communications · doi:10.1038/s41467-024-50698-y · Cited by 86
Structure and dynamics in drug discovery
Wei, H., & McCammon, J. A. (2024) · npj Drug Discovery. · doi:10.1038/s44386-024-00001-2 · Cited by 101
Roche, R., Moussad, B., Shuvo, M. H., Tarafder, S., & Bhattacharya, D. (2024) · Nucleic Acids Research · doi:10.1093/nar/gkae039 · Cited by 85
Kaminski, K., Ludwiczak, J., Pawlicki, K., Alva, V., & Dunin-Horkawicz, S. (2023) · Bioinformatics · doi:10.1093/bioinformatics/btad579 · Cited by 84
Song, Y., Yuan, Q., Chen, S., Zeng, Y., Zhao, H., & Yang, Y. (2024) · Nature Communications · doi:10.1038/s41467-024-52533-w · Cited by 89
Enhancing efficiency of protein language models with minimal wet-lab data through few-shot learning
Zhou, Z., Zhang, L., Yu, Y., Wu, B., Li, M., Hong, L., & Tan, P. (2024) · Nature Communications · doi:10.1038/s41467-024-49798-6 · Cited by 93
Bushuiev, R., Bushuiev, A., Samusevich, R., Brungs, C., Šivic, J., & Pluskal, T. (2025) · Nature Biotechnology · doi:10.1038/s41587-025-02663-3 · Cited by 105
Lu, Z., Song, G., Zhu, H., Lei, C., Sun, X., Wang, K., Qin, L., Chen, Y., Tang, J. P., & Li, M. (2025) · Nature Communications · doi:10.1038/s41467-025-57828-0 · Cited by 99
Nicheformer: a foundation model for single-cell and spatial omics
Tejada-Lapuerta, A., Schaar, A. C., Gutgesell, R., Palla, G., Halle, L., Minaeva, M., Vornholz, L., Dony, L., Drummer, F., Richter, T., et al. (2025) · Nature Methods · doi:10.1038/s41592-025-02814-z · Cited by 100
Prediction of Klebsiella phage-host specificity at the strain level
Boeckaerts, D., Stock, M., Ferriol-González, C., Oteo, J., Sanjuán, R., Domingo‐Calap, P., Bernard De Baets, C., & Briers, Y. (2024) · Nature Communications · doi:10.1038/s41467-024-48675-6 · Cited by 81
Advancing microbial production through artificial intelligence-aided biology
Gong, X., Zhang, J., Gan, Q., Teng, Y., Hou, J., Lyu, Y., Liu, Z., Wu, Z., Dai, R., Zou, Y., et al. (2024) · Biotechnology Advances · doi:10.1016/j.biotechadv.2024.108399 · Cited by 85
A virally encoded tRNA neutralizes the PARIS antiviral defence system
Burman, N., Belukhina, S., Depardieu, F., Wilkinson, R. A., Skutel, M. A., Santiago‐Frangos, A., Graham, A. B., Livenskyi, A., Chechenina, A., Morozova, N. E., et al. (2024) · Nature · doi:10.1038/s41586-024-07874-3 · Cited by 88
Bartolec, T. K., Vázquez-Campos, X., Norman, A. R., Luong, C., Johnson, M., Payne, R. J., Wilkins, M. R., Mackay, J. P., & Low, J. K. K. (2023) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2219418120 · Cited by 83
Codon language embeddings provide strong signals for use in protein engineering
Outeiral, C., & Deane, C. M. (2024) · Nature Machine Intelligence · doi:10.1038/s42256-024-00791-0 · Cited by 90
Advanced polishing methods for atomic-scale surfaces: A review
Luo, Z., Zhang, Z., Zhao, F., et al. (2024) · Materials Today Sustainability · doi:10.1016/j.mtsust.2024.100841 · Cited by 79
Wee, J., & Wei, G. (2024) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.4c00976 · Cited by 103
Liu, Y., & Tian, B. (2023) · Briefings in Bioinformatics · doi:10.1093/bib/bbad488 · Cited by 78
The future of rapid and automated single-cell data analysis using reference mapping
Lotfollahi, M., Hao, Y., Theis, F. J., & Satija, R. (2024) · Cell · doi:10.1016/j.cell.2024.03.009 · Cited by 81
Geometry-complete diffusion for 3D molecule generation and optimization
Morehead, A., & Cheng, J. (2024) · Communications Chemistry · doi:10.1038/s42004-024-01233-z · Cited by 102
He, Y., Zhou, X., Chang, C. H., Chen, G., Liu, W., Li, G., Fan, X., Sun, M., Miao, C., Huang, Q., et al. (2024) · Molecular Cell · doi:10.1016/j.molcel.2024.01.021 · Cited by 79
InterPLM: discovering interpretable features in protein language models via sparse autoencoders
Simon, E. P., & Zou, J. (2025) · Nature Methods · doi:10.1038/s41592-025-02836-7 · Cited by 89
AlphaFold two years on: Validation and impact
Kovalevskiy, O. V., Mateos-Garcia, J. C., & Tunyasuvunakool, K. (2024) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2315002121 · Cited by 92
Predicting the subcellular location of prokaryotic proteins with DeepLocPro
Moreno, J., Nielsen, H., Winther, O., & Teufel, F. (2024) · Bioinformatics · doi:10.1093/bioinformatics/btae677 · Cited by 76
Prediction of hemolytic peptides and their hemolytic concentration
Rathore, A. S., Kumar, N., Choudhury, S., Mehta, N. K., & Raghava, G. P. S. (2025) · Communications Biology · doi:10.1038/s42003-025-07615-w · Cited by 87
AlphaFold Protein Structure Database 2025: a redesigned interface and updated structural coverage
Bertoni, D., Tsenkov, M. I., Magaña, P., Nair, S. S., Pidruchna, I., Afonso, M. Q. L., Midlik, A., Paramval, U., Lawal, D. K., Tanweer, A., et al. (2025) · Nucleic Acids Research · doi:10.1093/nar/gkaf1226 · Cited by 101
Liu, J., Guan, S., Zou, Q., Wu, H., Tiwari, P., & Ding, Y. (2023) · Knowledge-Based Systems · doi:10.1016/j.knosys.2023.111329 · Cited by 77
Language models for biological research: a primer
Simon, E. P., Swanson, K., & Zou, J. (2024) · Nature Methods · doi:10.1038/s41592-024-02354-y · Cited by 85
chiLife: An open-source Python package for in silico spin labeling and integrative protein modeling
Tessmer, M. H., & Stoll, S. (2023) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1010834 · Cited by 82
Proteins with alternative folds reveal blind spots in AlphaFold-based protein structure prediction
Chakravarty, D., Lee, M., & Porter, L. L. (2025) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2024.102973 · Cited by 76
Ojala, T., Häkkinen, A.-E., Kankuri, E. M., & Kankainen, M. (2023) · Trends in Genetics · doi:10.1016/j.tig.2023.05.004 · Cited by 81
Artificial intelligence for science in quantum, atomistic, and continuum systems
Zhang, X., Wang, L., Helwig, J., Luo, Y., Fu, C., Xie, Y., Liu, M., Lin, Y., Zhao, X., Yan, K., et al. (2025) · Foundations and Trends® in Machine Learning · doi:10.1561/2200000115 · Cited by 214
Diffusion models in protein structure and docking
Yim, J., Stärk, H., Corso, G., Jing, B., Barzilay, R., & Jaakkola, T. (2024) · Wiley Interdisciplinary Reviews Computational Molecular Science · doi:10.1002/wcms.1711 · Cited by 87
Cheng, P., Mao, C., Tang, J., Yang, S., Cheng, Y., Wang, W., Gu, Q., Han, W., Chen, H., Li, S., et al. (2024) · Cell Research · doi:10.1038/s41422-024-00989-2 · Cited by 76
Metadata
- Authors
- Lin, Z., Akin, H., Rao, R., Hie, B., Zhu, Z., Lu, W., Smetanin, N., Verkuil, R., Kabeli, O., Shmueli, Y., dos Santos Costa, A., Fazel-Zarandi, M., Sercu, T., Candido, S., & Rives, A.
- Affiliations
- New York University (US)
- Stanford University (US)
- Corresponding author
- Rives, Alexander
- ORCID
- Venue
- Science
- Publisher
- openRxiv
- Published
- 2022-07-21
- Volume
- 379
- Issue
- 6637
- Pages
- 1123-1130
- ISSN
- 0036-8075 (ISSN-L)
- OpenAlex
- W4327550249
- Cited by
- 5,990
- References
- 71
- Citations by year
- 2026: 2468 · 2025: 1837 · 2024: 1170 · 2023: 485 · 2022: 23
- FWCI
- 460.33
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- Science (publishedVersion, cc-by)
- bioRxiv (Cold Spring Harbor Laboratory) (submittedVersion, cc-by-nc-nd)
- PubMed (publishedVersion)
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Structure and Dynamics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Machine Learning in Bioinformatics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Structural Biology › Enzyme Structure and Function
- Keywords
- protein structure prediction, large language models, protein evolution, atomic-scale structure, protein folding, structural diversity
- MeSH terms
- Machine Learning, Amino Acid Sequence, Protein Conformation, Proteins, Evolution, Molecular, Sequence Analysis, Protein
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
