The DOI matches its Crossref record; 77 of the 77 references it cites were found in a registry or an index; 1 of the 23 sources its claims cite were checked for support.
Large language models generate functional protein sequences across diverse families
Open accessAbstract
Deep-learning language models have shown promise in various biotechnological applications, including protein design and engineering. Here we describe ProGen, a language model that can generate protein sequences with a predictable function across large protein families, akin to generating grammatically and semantically correct natural language sentences on diverse topics. The model was trained on 280 million protein sequences from >19,000 families and is augmented with control tags specifying protein properties. ProGen can be further fine-tuned to curated sequences and tags to improve controllable generation performance of proteins from families with sufficient homologous samples. Artificial proteins fine-tuned to five distinct lysozyme families showed similar catalytic efficiencies as natural lysozymes, with sequence identity to natural proteins as low as 31.4%. ProGen is readily adapted to diverse protein families, as we demonstrate with chorismate mutase and malate dehydrogenase.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
32 Own result · 14 Own interpretation · 11 Not checked yet · 8 Own method · 4 No basis given · 1 Couldn’t confirm · accuracy not yet measured
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“As in the crystal structure of L056, the predicted artificial structures roughly match known structures found in nature, Fig S11, including for low identity (<40%) artificial sequences.”
§ Large language models generate functional protein sequences across diverse families > Abstract
What the sources say
- Could not verifycites [14] Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021). sourceExperimental — accuracy not yet measured
The Abstract and Main section describe AlphaFold’s structure-prediction accuracy, including cases without a known similar structure. Our copy does not address whether predicted structures of functional artificial sequences resemble natural structures, including at below 40% sequence identity (searched: artificial sequences, L056, and 40% identity). (LLM, conf=0.77)
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Could not verify
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
“Our protein language model is trained by simply learning to predict the next amino acid for over 280 million protein sequences from thousands of protein families, without explicit biophysical modeling.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“After training, ProGen can be prompted to generate full-length protein sequences for any protein family from scratch, with a varying degree of similarity to natural proteins.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“Our artificial lysozymes span the sequence landscape of natural lysozymes, Fig. 2a, across five families which contain diverse protein folds, active site architectures, and enzymatic mechanisms49,50.”
§ Large language models generate functional protein sequences across diverse families > Abstract
What the sources say
- Not checked yetcites [49] Baase, W. A., Liu, L., Tronrud, D. E. & Matthews, B. W. Lessons from the lysozyme of phage T4. Protein Sci. 19, 631–641 (2010). source
- Not checked yetcites [50] Kuroki, R., Weaver, L. H. & Matthews, B. W. A covalent enzyme–substrate intermediate with saccharide distortion in a mutant T4 lysozyme. Science 262, 2030–2033 (1993). source
plainn = 5- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption: The five lysozyme families chosen because they lack disulfide bonds can be expressed in high-throughput cell-free systems.
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
“they demonstrate similar residue position entropies when forming separate multiple sequence alignments of natural and artificial proteins within each family, Fig. 2c.”
§ Large language models generate functional protein sequences across diverse families > Abstract
boostedn = 5- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why 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
“In the positive control set of 100 natural proteins, 72% were well-expressed based on chromatography peaks and band visualization.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why 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
“Our artificial proteins match activity levels of natural proteins even at lower levels of sequence identity to any known natural protein, (Fig 3b, Fig. S5).”
§ Large language models generate functional protein sequences across diverse families > Abstract
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why 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
“The representative natural proteins exhibited similar levels of functionality with 59% (53/90) of total proteins considered functional.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“None of the bmDCA artificial proteins exhibited a detectable level of functionality, Fig. S10”
§ Large language models generate functional protein sequences across diverse families > Abstract
What the sources say
- Not checked yetcites [7] Russ, W. P. et al. An evolution-based model for designing chorismate mutase enzymes. Science 369, 440–445 (2020). source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why 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
“Generated proteins exhibit similar conservation patterns to natural sequence libraries (Fig. 4ad).”
§ Large language models generate functional protein sequences across diverse families > Abstract
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- It may state two separate claims as one. 29
- Probably not a substantial claim. 16
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
“ProGen is a 1.2 billion-parameter neural network trained using a publicly available dataset of 280 million protein sequences.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“we could readily measure the KM values for the purified apparent lower molecular weight species, where both L056 and L070 were highly active and had comparable Michaelis-Menten parameters to HEWL (Fig. 3e).”
§ Large language models generate functional protein sequences across diverse families > Abstract
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“Among our artificial proteins, 73% (66/90) were functional and exhibited high levels of functionality across families, Fig. 3c.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“Notably a small number of proteins, both within the natural and artificial proteins, were within an order of magnitude of HEWL, which was substantially more active than all negative controls.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“Two proteins, L056 (max ID 69.6%) and L070 (max ID 89.2%) expressed well and incurred bactericidal activities towards the E. coli BL21(DE3) strain used during overnight induction at 16C.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“The ProGen-generated proteins express equally well (72/100 total) across all bins of sequence identity to any known natural protein (maxID 40–90%), Fig 2e.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“Despite having comparable enzymatic activities, L070 and L056 only share 17.9% sequence ID.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“We also found that there is no bias to location or structural element to the mutations that diverge L056 and L070 from their respective nearest sequence homolog in nature. Differing residues are instead uniformly distributed.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“The KM values of both monomers were too weak to be measured using a heterogeneous, fluorescein-labeled Micrococcus lysodeikticus cell wall substrate (Molecular Probes EnzChek Lysozyme kit); however, both monomers were active using a pseudo-first order kinetic assay (Fig. S7).”
§ Large language models generate functional protein sequences across diverse families > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“Some mutations are even found within the active site cleft and within regions that influence conformational state (for L056).”
§ Large language models generate functional protein sequences across diverse families > Abstract
plainSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“We determined a 2.5 Å resolution crystal of L056 (see Methods, Figure 3e). The global fold was similar to predictions, with a Cα RMSD of 2.9 Å from the backbone structure predicted by trRosetta and 2.3 Å RMSD from a WT T4 Lysozyme structure59,60.”
§ Large language models generate functional protein sequences across diverse families > Abstract
What the sources say
plainSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“bmDCA was unable to fit three out of the five lysozyme families, and exhibited 60% detectable expression (30/50 proteins) for the remaining two protein families.”
§ Large language models generate functional protein sequences across diverse families > Abstract
What the sources say
- Not checked yetcites [7] Russ, W. P. et al. An evolution-based model for designing chorismate mutase enzymes. Science 369, 440–445 (2020). source
plainn = 5- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“Alignment with a structure featuring a covalently trapped substrate (PDB 148L) reveals that the active site cleft is well formed with the key catalytic residue Glu15 (Glu 11 in T4L) and key substrate binding residue Thr30 (Thr26 in T4L) correctly positioned.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plainSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“These likelihoods showed a significant drop in AUC of 0.18 for CM (p<0.0001, two-tailed test, n=1617) and 0.08 for MDH (p<0.1, two-tailed test, n=56), as compared to fine-tuned ProGen when predicting if an artificial sequence should function.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plaindirectn = 1,617 ·- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“On MDH function data, ProGen log-likelihoods had an AUC of 0.94 (Fig. 4f), which was better than ProteinGAN discriminator scores, with an AUC of 0.87 (p<0.1, two-tailed test, n=56).”
§ Large language models generate functional protein sequences across diverse families > Abstract
What the sources say
- Not checked yetcites [66] Vig, J. et al. BERTology meets biology: Interpreting attention in protein language models. In International Conference on Learning Representations (ICLR, 2020). source
plainn = 56 ·- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“On CM function data, ProGen log-likelihoods had an area under the curve (AUC) of 0.85, significantly better (p<0.0001, two-tailed test, n=1617) than bmDCA, which had an AUC of 0.78 (Fig. 4c).”
§ Large language models generate functional protein sequences across diverse families > Abstract
What the sources say
- Not checked yetcites [7] Russ, W. P. et al. An evolution-based model for designing chorismate mutase enzymes. Science 369, 440–445 (2020). source
plainn = 1,617- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“We purified 6 highly expressed proteins and found that they were all active, but with significantly lower Michaelis-Menten activities than HEWL or the previously generated artificial proteins L056 and L070 (Fig 3f).”
§ Large language models generate functional protein sequences across diverse families > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“Of the selected sequences, 78 out of the 89 (88%) expressed well and 24 out of the 78 (31%) were soluble (Supplemental Figure 8).”
§ Large language models generate functional protein sequences across diverse families > Abstract
plainn = 2- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“L013 and L038 expressed highly to inclusion bodies.”
§ Large language models generate functional protein sequences across diverse families > Methods > E. coli Expression of Lysozyme Variants
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
“For lowID lysozyme A5 the above protocol was altered slightly to accommodate lower catalytic activity of these variants: reaction volumes were increased to 20 μL, plate was covered with an optically transparent seal (Microseal ® ‘B’ seal; BioRad) to mitigate sample evaporation, fluorescent reads were taken every 5 mins for 16 hours with 5 s of linear plate shaking prior to each measurement to min”
§ Large language models generate functional protein sequences across diverse families > Methods > Michaelis-Menten kinetics of lysozyme variants using fluorescein-labeled M. lysodeikticus cell wall
plaindirectSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It 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
“The MDH functional labels are binary, so no thresholding was needed to compute AUCs.”
§ Large language models generate functional protein sequences across diverse families > Methods > Evaluating ProGen on Other Protein Systems
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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 resultant data were not described by a single exponential model but were described by a double exponential model (Eq. 6), likely due to the heterogeneity of the substrate, and all data were fit in IgorPro7.”
§ Large language models generate functional protein sequences across diverse families > Methods > Lysozyme kcat/KM extrapolation from pseudo first-order kinetic data
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
“For induced cultures of L056 and L070, significant turbidity was observed in the spent media after cells were pelleted at 3,500 r.c.f. for 30 mins at 4°C.”
§ Large language models generate functional protein sequences across diverse families > Methods > E. coli Expression of Lysozyme Variants
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
“Two peaks were resolved for each variant and that harbored lysozyme activity against the fluorescein-labeled M. Lysodeikticus cell wall substrate (EnzChek kit; ThermoFisher).”
- Also stated in: § Purification of L056 and L070 from spent media · plain
“Two peaks were resolved for each variant and that harbored lysozyme activity against the fluorescein-labeled M. Lysodeikticus cell wall substrate (EnzChek kit; ThermoFisher) that corresponded to peaks observed in the first pool purification.”
§ Large language models generate functional protein sequences across diverse families > Methods > Purification of L056 and L070 from spent media
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Also stated in: § Purification of L056 and L070 from spent media · plain
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
“Significant translational non-crystallographic symmetry and differences with the search model resulted in maps that were initially hard to interpret.”
§ Large language models generate functional protein sequences across diverse families > Methods > Crystallization and structure determination of L056
plaindirectSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
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
“These results indicate that ProGen generates protein sequences that not only can express well but also maintain enzymatic function for diverse sequence landscapes across protein families.”
§ Large language models generate functional protein sequences across diverse families > Abstract
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Rests on an assumption: Well-expression and solubility in the cell-free system indicate that the artificial proteins are structurally well-folded.
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
“Taken together, L056 and L070 harbor potent catalytic activity and bactericidal capabilities that are comparable to HEWL, while diverging from their nearest known natural sequence by 53 and 18 amino acids respectively.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Rests on an assumption: HEWL is a suitable reference for native lysozyme catalytic activity.
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 perplexity on a held-out test set was monitored and did not exceed training set perplexity throughout model training.”
§ Large language models generate functional protein sequences across diverse families > Methods > ProGen Training
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
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
“While we do not expect our language model to generate proteins that belong to a completely different distribution or domain (for example creating a new fold that catalyzes an unnatural reaction), it can significantly expand the space of protein sequences from those sampled by evolution.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Rests on an assumption: The 280 million-sequence training set is diverse enough to learn a universal protein representation transferable across families.
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 aggregated dataset was split into a training set of size 280 million and two test sets, an out-of-distribution test set (OOD-test) of size 100k from 20 protein families and a randomly sampled in-domain test set (ID-test) of size 1 million, that were held-out for training and used for evaluation.”
§ Large language models generate functional protein sequences across diverse families > Methods > Training Data Curation
plainn = 20- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“In sum, ProGen’s model likelihoods are better aligned with experimentally-measured assay data on two diverse protein datasets—chorismate mutase and malate dehydrogenase—than the sequence generation methods from original studies specifically tailored for these families.”
§ Large language models generate functional protein sequences across diverse families > Abstract
What the sources say
- Not checked yetcites [7] Russ, W. P. et al. An evolution-based model for designing chorismate mutase enzymes. Science 369, 440–445 (2020). source
- Not checked yetcites [66] Vig, J. et al. BERTology meets biology: Interpreting attention in protein language models. In International Conference on Learning Representations (ICLR, 2020). source
plaincorrelationaln = 1,617- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It 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
“These results indicate that both components of our training strategy—initial training on the universal sequence dataset and fine-tuning on the protein family of interest—contribute significantly to final model performance.”
§ Large language models generate functional protein sequences across diverse families > Abstract
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a 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
“For robustness, we trained three discriminators using different random seeds.”
§ Large language models generate functional protein sequences across diverse families > Methods > Lysozyme Sequence Selection
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
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
“Proteins that generated fluorescence one standard deviation above the maximum fluorescence of any negative control were considered functional.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
“The resultant data were not described by a single exponential model but were described by a double exponential model (Eq. 6), likely due to the heterogeneity of the substrate, and all data were fit in IgorPro7.”
§ Large language models generate functional protein sequences across diverse families > Methods > Lysozyme kcat/KM extrapolation from pseudo first-order kinetic data
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Rests on an assumption: The average molecular weight of the fluorescein-labeled M. lysodeikticus cell wall substrate is unknown and the substrate is heterogeneous.
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 also ensured that no two generated samples share over 80% identity with each other.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
“After measuring the AUC of each model for each dataset, we used bootstrapping to compute the statistical significance of the difference in AUC of fine-tuned ProGen vs. the reference method (bmDCA and ProGen ablations for CM, ProteinGAN and ProGen ablations for MDH).”
§ Large language models generate functional protein sequences across diverse families > Methods > Evaluating ProGen on Other Protein Systems
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
“The fine-tuning dataset was clustered at 80% sequence identity and 10% of the clusters were held-out as a validation set for hyper-parameter optimization and stopping criteria.”
§ Large language models generate functional protein sequences across diverse families > Methods > Lysozyme Generation
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“We also applied thresholding at 0.42 norm relative enrichment (r.e.) to obtain binary labels for CM function, which roughly corresponds to the cut-off point between two modes that exist in CM function data, to be used for ROC curves, following the original study7.”
§ Large language models generate functional protein sequences across diverse families > Methods > Evaluating ProGen on Other Protein Systems
What the sources say
- Not checked yetcites [7] Russ, W. P. et al. An evolution-based model for designing chorismate mutase enzymes. Science 369, 440–445 (2020). source
hedgedcorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
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 performed structure prediction without an MSA as input so as to not heavily bias the structure prediction toward a known natural mode.”
§ Large language models generate functional protein sequences across diverse families > Abstract
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim.
The paper relies on this without support.
The five lysozyme families chosen because they lack disulfide bonds can be expressed in high-throughput cell-free systems.
“Within the lysozyme clan, we chose to work with five families that lack disulfide bonds as to be amenable to high-throughput cell-free expression: Phage lysozyme, Pesticin, Glucosaminidase, Glycoside hydrolase, and Transglycosylase.”
The paper relies on this without support.
Well-expression and solubility in the cell-free system indicate that the artificial proteins are structurally well-folded.
“These results indicate that ProGen can generate artificial proteins that are structurally well-folded for proper expression as compared to a batch of natural proteins”
The paper relies on this without support.
The 280 million-sequence training set is diverse enough to learn a universal protein representation transferable across families.
“Trained in this unsupervised manner from a large, varied protein sequence database, ProGen learns a universal, domain-independent representation of proteins that subsumes local and global structure motifs”
The paper relies on this without support.
HEWL is a suitable reference for native lysozyme catalytic activity.
“Hen egg white lysozyme (HEWL), a naturally-evolved exemplar protein, was measured as positive control”
The paper relies on this without support.
Model likelihoods for GANs are intractable.
“Since model likelihoods for GANs are intractable, we used discriminator scores corresponding to the probability at which the ProteinGAN discriminator predicted each sample was real as a ProteinGAN model score for each MDH sequence.”
The paper relies on this without support.
The average molecular weight of the fluorescein-labeled M. lysodeikticus cell wall substrate is unknown and the substrate is heterogeneous.
“(the average molecular weight of the fluorescein-labeled M. Lysodeikticus cell wall substrate was unknown and likely heterogeneous).”
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 source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“synthetic proteins optimized for specific families of lysozymes using ProGen have exhibited catalytic efficiencies comparable to their naturally occurring counterparts with low sequence identity.”
From Large language models (LLMs): survey, technical frameworks, and future challenges (2024) · cites this paper as Madani et al. 2023
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“AI models have already been used to design new CRISPR enzymes,53 functional proteins,128 and even entire prokaryotic genomes.51”
From How to build the virtual cell with artificial intelligence: Priorities and opportunities (2024) · cites this paper as [128]
The cited source says what the claim attributes to it.
In the citing paper
“The success of LLMs has revolutionized AI learning in natural language and image processing [75] and has been extended into the learning of biological sequence datasets [80].”
From Challenges in AI-driven Biomedical Multimodal Data Fusion and Analysis (2025) · cites this paper as [80]
The cited source says what the claim attributes to it.
In the citing paper
“the ProGen (Madani et al. 2023) language model employs deep learning techniques to produce protein sequences with predefined functionalities.”
From Large language models (LLMs): survey, technical frameworks, and future challenges (2024) · cites this paper as Madani et al. 2023
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.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Large language models generate functional protein sequences across diverse families > Abstract
“a state-of-the-art transformer-based conditional language model trained only with evolutionary sequence data”
§ Large language models generate functional protein sequences across diverse families > Methods > ProGen Training
“The model was initialized with pretrained weights of CTRL28, which was trained on an English language corpus.”
Citations
Cites
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
LeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
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
Gene Ontology: tool for the unification of biology
Ashburner, M., Ball, C. A., Blake, J. A., Botstein, D., Butler, H., Cherry, J. M., Davis, A. P., Dolinski, K., Dwight, S. S., Eppig, J. T., et al. (2000) · Nature Genetics · doi:10.1038/75556 · Cited by 43,206
Features and development of Coot
Emsley, P., Lohkamp, B., Scott, W. G., & Cowtan, K. (2010) · Acta Crystallographica Section D Biological Crystallography · doi:10.1107/s0907444910007493 · Cited by 29,915
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
Phaser crystallographic software
McCoy, A. J., Grosse-Kunstleve, R. W., Adams, P. D., Winn, M. D., Storoni, L. C., & Read, R. J. (2007) · Journal of Applied Crystallography · doi:10.1107/s0021889807021206 · Cited by 20,612
Kabsch, W. (2010) · Acta Crystallographica Section D Biological Crystallography · doi:10.1107/s0907444909047337 · Cited by 16,463
Deep Contextualized Word Representations
Peters, M., Neumann, M., Iyyer, M., Gardner, M., Clark, C., Lee, K., & Zettlemoyer, L. (2018) · Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers) · doi:10.18653/v1/n18-1202 · Cited by 11,983
Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., & Liu, P. J. (2019) · arXiv · doi:10.48550/arxiv.1910.10683 · Cited by 10,929
Research works printed but not linked: 13
- Vaswani, A. et al. Attention is all you need. In 31st Conference on Neural Information Processing Systems (NIPS, 2017).
- Devlin, J., Chang, M.-W., Lee, K. & Toutanova, K. BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT, 2019).
- Zellers, R. et al. Defending against neural fake news. In 33rd Conference on Neural Information Processing Systems (NeurIPS, 2019).
- Radford, A., Narasimhan, K., Salimans, T. & Sutskever, I. Improving language understanding by generative pre-training. Preprint at https://s3-us-west-2.amazonaws.com/openai-assets/research-covers/language-unsupervised/language_understanding_paper.pdf (2018).
- Dathathri, S. et al. Plug and play language models: a simple approach to controlled text generation. In 8th International Conference on Learning Representations (ICLR, 2020).
- Pettit, L. D. The IUPAC stability constants database. Chem. Int. 28, 14–15 (2006).
- Bengio, Y., Ducharme, R., Vincent, P. & Janvin, C. A neural probabilistic language model. J. Mach. Learn. Res. 3, 1137–1155 (2003).
- Goyal, K., Dyer, C. & Berg-Kirkpatrick, T. Exposing the implicit energy networks behind masked language models via metropolis–hastings. In 10th International Conference on Learning Representations (ICLR, 2022).
- Alley, E., Khimulya, G., Biswas, S., AlQuraishi, M. & Church, G. M. Unified rational protein engineering with sequence-based deep representation learning. Nat. Methods 16, 1315–1322 (2019). · DOI · Paper page
- Pascanu, R., Mikolov, T. & Bengio, Y. On the difficulty of training recurrent neural networks. In Proc. 30th International Conference on Machine Learning (eds. Dasgupta, S. & McAllester, D.) 1310–1318 (PMLR, 2013).
- Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I. & Salakhutdinov, R. Dropout: a simple way to prevent neural networks from overfitting. J. Mach. Learn. Res. 15, 1929–1958 (2014).
- Holtzman, A., Buys, J., Du, L., Forbes, M. & Choi, Y. The curious case of neural text degeneration. In 8th International Conference on Learning Representations (ICLR, 2020).
- Goodfellow, I. J. et al. Generative adversarial networks. In 28th Conference on Neural Information Processing Systems (NIPS, 2014).
Cited by
Sorted by citations
Items 101–200 of 1,012
The past, present and future of de novo protein design
Yang, W., Wang, S., Lee, G. R., Zhang, J. Z., Courbet, A., Juergens, D., Wang, X., Schlichthaerle, T., Abedi, M. H., Ragotte, R. J., et al. (2026) · Nature · doi:10.1038/s41586-026-10328-7 · Cited by 72
“Generative AI methods 116 such as RFdiffusion 26 and gradient-based hallucination approaches such as Bindcraft have been used to robustly generate binders against many protein targets 117 ; other methods include Chroma 27 , ProGen 52 and PXDesign 118 .”
Large language models for drug discovery and development
Zheng, Y., Koh, H. Y., Ju, J., Yang, M., May, L. T., Webb, G. I., Li, L., Pan, S., & Church, G. M. (2025) · Patterns · doi:10.1016/j.patter.2025.101346 · Cited by 121
“94,95 Unconstrained generation, using specialized LLMs like ProtGPT2 91 and ProGen, 96 has demonstrated success in generating novel sequences that resemble natural proteins, while also exploring uncharted protein space.”
Electromagnetic metamaterial agent
Hu, S., Li, M., Xu, J., Zhang, H., Zhang, S., Cui, T. J., del Hougne, P., & Li, L. (2025) · Light Science & Applications · doi:10.1038/s41377-024-01678-w · Cited by 50
“LFMs have successfully been applied to areas including task planning 38,39 , computer code generation 40 , protein design 41 , and information extraction from materials science literature 42 .”
Large language models for science and medicine
Telenti, A., Auli, M., Hie, B., Maher, C., Saria, S., & Ioannidis, J. P. A. (2024) · European Journal of Clinical Investigation · doi:10.1111/eci.14183 · Cited by 62
“18 Language models can also be trained on sequences of nucleic acids that form DNA and RNA molecules.”
BackgroundGenomic 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
“Starting with a given sequence fragment (referred to as prompts [25], or con-trol tags [26]), the language model can predict the next token recursively and generate a whole new sequence. pLMs have been shown to be powerful tools for protein design [26–29].”
Protein Large Language Models: A Comprehensive SurveyPreprint
Xiao, Y., Zhao, W., Zhang, J., Jin, Y., Zhang, H., Ren, Z., Sun, R., Wang, H., Wan, G., Pan, L. Y., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2502.17504 · Cited by 39
“Pretraining Dataset: AlphaFoldDB (Tunyasuvunakool et al., 2021) Benchmark Dataset: ProteinGym (Notin et al., 2023), ProteinLMBench (Shen et al., 2024a) Method LLM Methods for Protein Sequence Models UniRep (Alley et al., 2019), SSA (Bepler and Berger, 2019), MuPIPR (Zhou et al., 2020), ProtTrans(Elnaggar et al., 2021)…”
GenerRNA: A generative pre-trained language model for de novo RNA design
Zhao, Y., Oono, K., Takizawa, H., & Kotera, M. (2024) · PLoS ONE · doi:10.1371/journal.pone.0310814 · Cited by 50
“To address this, there have been recent efforts to integrate deep learning approaches into RNA generation, such as RNAGEN [7], which utilizes WGAN-GP [8], a variant of the generative adversarial network (GAN) [9] architecture, to generate new sequences with similar characteristics to training…”
MethodAI for targeted polypharmacology: The next frontier in drug discovery
Cichońska, A., Ravikumar, B., & Rahman, R. (2024) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2023.102771 · Cited by 68
Lab-in-the-loop therapeutic antibody design with deep learningPreprint
Frey, N. C., Hötzel, I., Stanton, S., Kelly, R. L., Alberstein, R. G., Makowski, E. K., Martinkus, K., Berenberg, D., Bevers, J. J., Bryson, T., et al. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.02.19.639050 · Cited by 33
Conditional language models enable the efficient design of proficient enzymesPreprint
Munsamy, G., Illanes-Vicioso, R., Fruncillo, S., Nakou, I. T., Lindner, S., Ayres, G., Sheehan, L. S., Moss, S. J., Eckhard, U., Lorenz, P., et al. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.05.03.592223 · Cited by 59
“…other generative models have provided artificial variants for several enzyme families 19,20 , including artificial TEM-1 β-lactamases 21 , lysozymes 22 , luciferases 2 , malate dehydrogenases 23 , superoxide dismutases 24 , chorismate mutases 25 , and CRISPR-Cas genome editors 26 in the last three…”
BackgroundMethodInfluentialProteinGPT: Multimodal LLM for Protein Property Prediction and Structure UnderstandingPreprint
Xiao, Y., Sun, E., Jin, Y., Wang, Q., & Wang, W. W. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2408.11363 · Cited by 50
“Some research [24, 25] treat protein sequences as a biological language and utilize the Transformer architecture [26] to model amino acid interactions and predict sequences using large sequence databases.”
MethodProtT3: Protein-to-Text Generation for Text-based Protein Understanding
Liu, Z., Zhang, A., Fei, H., Zhang, E., Wang, X., Kawaguchi, K., & Chua, T.-S. (2024) · Annual Meeting of the Association for Computational Linguistics · doi:10.48550/arxiv.2405.12564 · Cited by 51
“PLMs are transformer-based LMs pretrained on large corpora of protein sequences for protein understanding and protein generation (Lin et al., 2022; Madani et al., 2023; Nijkamp et al., 2022; Chen et al., 2023; El-naggar et al., 2021; Rives et al., 2021a; Meier et al., 2021).”
BackgroundMethodAlbanese, K. I., Barbe, S., Tagami, S., Woolfson, D. N., & Schiex, T. (2025) · Nature Reviews Methods Primers · doi:10.1038/s43586-025-00383-1 · Cited by 104
Structure-aware machine learning strategies for antimicrobial peptide discovery
Aguilera‐Puga, M. D. C., & Plisson, F. (2024) · Scientific Reports · doi:10.1038/s41598-024-62419-y · Cited by 52
Zeng, W., Dou, Y., Pan, L., Xu, L., & Peng, S. (2024) · Nature Communications · doi:10.1038/s41467-024-52293-7 · Cited by 49
Are protein language models the new universal key?
Weißenow, K., & Rost, B. (2025) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2025.102997 · Cited by 54
Multi-CGAN: Deep Generative Model-Based Multiproperty Antimicrobial Peptide Design
Yu, H., Wang, R.-H., Qiao, J.-B., & Wei, L.-Y. (2023) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.3c01881 · Cited by 49
ProteinNPT: Improving Protein Property Prediction and Design with Non-Parametric Transformers
Notin, P., Weitzman, R., Marks, D. S., & Gal, Y. (2023) · bioRxiv · doi:10.1101/2023.12.06.570473 · Cited by 49
“…been used to sample natural-like proteins with no bias towards properties of interest [Repecka et al., 2021, Sgarbossa et al., 2023], or conditioned on broad taxonomic labels characterizing full protein families [Madani et al., 2023] as opposed to the sequence-specific properties we consider here.”
MethodGenerative Artificial Intelligence in Anatomic Pathology
Brodsky, V., Ullah, E., Bychkov, A., Song, A. H., Walk, E. E., Louis, P., Rasool, G., Singh, R. S., Mahmood, F., Bui, M. M., et al. (2025) · Archives of Pathology & Laboratory Medicine · doi:10.5858/arpa.2024-0215-ra · Cited by 48
Addressing epistasis in the design of protein function
Lipsh‐Sokolik, R., & Fleishman, S. J. (2024) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2314999121 · Cited by 55
Deep generative models for detector signature simulation: A taxonomic review
Hashemi, H., & Krause, C. (2024) · Reviews in Physics · doi:10.1016/j.revip.2024.100092 · Cited by 56
Generative artificial intelligence for enzyme design: Recent advances in models and applications
Wen, S., Zheng, W., Bornscheuer, U. T., & Wu, S. (2025) · Current Opinion in Green and Sustainable Chemistry · doi:10.1016/j.cogsc.2025.101010 · Cited by 47
Optical multilayer thin film structure inverse design: From optimization to deep learning
Ma, T., Ma, M., & Guo, L. J. (2025) · iScience · doi:10.1016/j.isci.2025.112222 · Cited by 44
“In both cases, one would need to collect a new dataset and retrain deep learning models.”
Can Large Language Models Understand Symbolic Graphics Programs?Preprint
Qiu, Z., Liu, W., Feng, H., Liu, Z., Xiao, T. Z., Collins, K. M., Tenenbaum, J. B., Weller, A., Black, M. J., & Schölkopf, B. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2408.08313 · Cited by 26
“LLMs have demonstrated growing potential in many applications, ranging from mathematical problem solving and theorem proving assistance [65, 114, 116, 15] to aiding biological discovery [67, 27, 94, 59].”
Modeling protein–ligand interactions for drug discovery in the era of deep learning
Wang, Y. X., Li, Y., Chen, J., & Lai, L. (2025) · Chemical Society Reviews · doi:10.1039/d5cs00415b · Cited by 53
Large language model for knowledge synthesis and AI-enhanced biomanufacturing
Li, W., Mao, Z., Xiao, Z., Liao, X., Koffas, M. A. G., Chen, Y., Ma, H., & Tang, Y. (2025) · Trends in biotechnology · doi:10.1016/j.tibtech.2025.02.008 · Cited by 49
Viral emergence and pandemic preparedness in a One Health framework
Sikkema, R. S., & Koopmans, M. (2025) · Nature Reviews Microbiology · doi:10.1038/s41579-025-01243-1 · Cited by 47
“Looking forward, the continuing expansion of metagenomic datasets combined with artificial intelligence approaches could have an important role in the analysis of metagenomic data and associated possible phenotype and zoonotic risk prediction 132,133 .”
Lin, A., Ye, J., Qi, C., Zhu, L., Mou, W., Gan, W., Zeng, D., Tang, B., Xiao, M., Chu, G., et al. (2025) · Briefings in Bioinformatics · doi:10.1093/bib/bbaf357 · Cited by 43
“ProGen, a next-generation large language model, generates protein sequences with predictable functions within extensive protein families, substantially enhancing the controllability and accuracy of homologous family protein sequence and function prediction [25].”
Lactylation of HADHA Promotes Sepsis-Induced Myocardial Depression.
Zhang, T.-N., Huang, X.-M., Li, L., Li, Y., Liu, Y.-P., Shi, X., Wu, Q.-J., Wen, R., Yang, Y.-H., Zhang, T., et al. (2025) · Circulation Research · doi:10.1161/circresaha.124.325708 · Cited by 43
Zhou, S., Liu, K., Li, D., Fu, C., Ning, Y., Ji, W., Liu, X., Xiao, B., & Wei, R. (2025) · Advanced Engineering Informatics · doi:10.1016/j.aei.2025.103142 · Cited by 49
CARE: a Benchmark Suite for the Classification and Retrieval of Enzymes
Yang, J., Mora, A., Liu, S., Wittmann, B. J., Anandkumar, A., Arnold, F. H., & Yue, Y.-S. (2024) · Neural Information Processing Systems · doi:10.48550/arxiv.2406.15669 · Cited by 43
“With the emergence of conditional generative models for protein sequences, it is also important to have high-throughput computational methods that can predict the function of generated sequences [80, 81, 82, 83, 84].”
BackgroundZhu, Y., Liu, Z., Liu, Y., Ji, Z., & Yu, D. (2024) · Briefings in Bioinformatics · doi:10.1093/bib/bbae040 · Cited by 49
“These methods are frequently employed through supervised deep neural networks in various bioinformatics tasks, including protein design [39, 40], function annotation [41, 42], structure prediction [43, 44] and ligand-binding prediction [45, 46].”
MethodLarge language models: a primer and gastroenterology applications
Shahab, O., Kurdi, B. E., Shaukat, A., Nadkarni, G. N., & Soroush, A. (2024) · Therapeutic Advances in Gastroenterology · doi:10.1177/17562848241227031 · Cited by 47
“66 An LLM-based model developed to predict a protein’s function from its sequence could accelerate the development of synthetic proteins for the treatment of inflammatory bowel disease and gastrointestinal cancers.”
Ni, B., Kaplan, D. L., & Buehler, M. J. (2024) · Science Advances · doi:10.1126/sciadv.adl4000 · Cited by 51
“In contrast, development of generative models aimed at end-to-end design from property of interest to protein sequence remain rare ( 60 ) .”
BackgroundMuir, D. F., Asper, G. P. R., Notin, P., Posner, J. A., Marks, D. S., Keiser, M. J., & Pinney, M. M. (2024) · doi:10.1101/2024.10.23.619915 · Cited by 42
“Furthermore, early PLM work demonstrated that the models learned an underlying representation of protein sequence that organized amino acids by biochemical property ( 34 ), and recent work has suggested their utility in optimizing protein function directly ( 35 , 36 ).”
Medium-sized protein language models perform well at transfer learning on realistic datasets
Vieira, L. C., Handojo, M. L., & Wilke, C. O. (2025) · Scientific Reports · doi:10.1038/s41598-025-05674-x · Cited by 41
“On the other hand, generative pre-trained transformer (GPT) models, such as ProGen 12 , are designed to generate novel protein sequences and have shown potential to create functional proteins.”
Jin, S., Zeng, Z., Xiong, X., Huang, B., Tang, L., Wang, H., Ma, X., Tang, X., Shao, G., Huang, X., et al. (2025) · Communications Biology · doi:10.1038/s42003-025-08282-7 · Cited by 43
“The fi rst type is sequence-based models, which aim to capture the biochemical constraints that characterize the proteins within the training set derived from extensive sequence data, with ProGen being a notable example 1 – 5 .”
AI-driven de novo enzyme design: Strategies, applications, and future prospects
Cui, X.-C., Zheng, Y., Liu, Y., Yuchi, Z., & Yuan, Y. (2025) · Biotechnology Advances · doi:10.1016/j.biotechadv.2025.108603 · Cited by 44
Roll the dice & look before you leap: Going beyond the creative limits of next-token predictionPreprint
Nagarajan, V., Wu, C. H., Ding, C., & Raghunathan, A. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2504.15266 · Cited by 21
“The exact form of coh will be defined in each algorithmic task but broadly, we are interested in scenarios where determining coherence requires a global understanding of the whole string.”
Optimizing enzyme thermostability by combining multiple mutations using protein language model
Bian, J., Tan, P., Nie, T., Hong, L. Y., & Yang, G. (2024) · mLife · doi:10.1002/mlf2.12151 · Cited by 43
Research on Bitter Peptides in the Field of Bioinformatics: A Comprehensive Review
Liu, S., Shi, T., Yu, J., Li, R., Lin, H., & Deng, K. (2024) · International Journal of Molecular Sciences · doi:10.3390/ijms25189844 · Cited by 42
“The latest de novo protein generative models, such as ProGen [200] and ProteinGAN [201], can produce entirely new proteins not yet found in nature, providing new ideas and insights for protein design [202–204].”
BackgroundProteinBench: A Holistic Evaluation of Protein Foundation ModelsPreprint
Ye, F., Zheng, Z., Xue, D., Shen, Y., Wang, L., Ma, Y., Wang, Y., Wang, X., Zhou, X., & Gu, Q. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2409.06744 · Cited by 31
“…by capturing the intricate mechanisms of proteins (Jumper et al., 2021; Abramson et al., 2024; Lin et al., 2023a; Watson et al., 2023b; Ingraham et al., 2023; Krishna et al., 2024; Shin et al., 2021; Madani et al., 2023; Alley et al., 2019; Wang et al., 2024b; Hayes et al., 2024; Hie et al., 2024).”
Deep Learning for Antimicrobial Peptides: Computational Models and Databases
Zhou, X., Liu, G., Cao, S., & Lv, J. Z. (2025) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.5c00006 · Cited by 37
Structure Language Models for Protein Conformation GenerationPreprint
Lu, J., Chen, X., Lu, S. Z., Shi, C., Guo, H., Bengio, Y., & Tang, J. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2410.18403 · Cited by 15
“On the other hand, auto-regressive protein language models, based on either recurrent neural networks (Al-ley et al., 2019), or Transformer including ProGen (Madani et al., 2023) and ProtGPT2 (Ferruz et al., 2022), are able to generate de novo sequences with input controlling tokens.”
Jiang, R., Yue, Z., Shang, L., Wang, D., & Wei, N. (2024) · Metabolic Engineering Communications · doi:10.1016/j.mec.2024.e00248 · Cited by 42
“…Liu et al., 2022; Mitchell and Mitchell, 1997), have been applied to protein function prediction (Yu et al., 2023), mutation identification (Hie et al., 2024; Lu et al., 2022; Shin et al., 2021), and protein design problems (Ferruz and H ¨ ocker, 2022; Ferruz et al., 2022; Madani et al., 2023).”
BackgroundMethodChen, T., Vure, P., Pulugurta, R., & Chatterjee, P. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.03.03.583201 · Cited by 31
“Concurrently, the development of protein language models (pLMs) such as ESM-2 (10), ProtT5 (11), ProGen (12), and ProtGPT2 (13), has significantly propelled our understanding and design capabilities for proteins.”
BackgroundPang, Y.-H., & Liu, B. (2024) · BMC Biology · doi:10.1186/s12915-023-01803-y · Cited by 38
Crowdsourced Protein Design: Lessons From the Adaptyv EGFR Binder CompetitionPreprint
Cotet, T., Krawczuk, I., Stocco, F., Ferruz, N., Gitter, A., Kurumida, Y., Machado, L. D. A., Paesani, F., Calia, C. N., Challacombe, C. A., et al. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.04.17.648362 · Cited by 32
transformative potential of Generative Artificial Intelligence (GenAI) in business
Caño-Marín, E. (2024) · ESIC MARKET Economic and Business Journal · doi:10.7200/esicm.55.333 · Cited by 53
Hourigan, D., Stefanovic, E., Hill, C., & Ross, R. P. (2024) · BMC Microbiology · doi:10.1186/s12866-024-03243-2 · Cited by 37
The promises of large language models for protein design and modeling
Valentini, G., Malchiodi, D., Gliozzo, J., Mesiti, M., Soto-Gomez, M., Cabri, A., Reese, J., Casiraghi, E., & Robinson, P. N. (2023) · Frontiers in Bioinformatics · doi:10.3389/fbinf.2023.1304099 · Cited by 40
“…prediction, remote homology detection (Rives et al., 2021; Brandes et al., 2022), de novo generation of functionally characterized proteins (Madani et al., 2023), design of antibodies that bind to speci fi c ligands (Hie et al., 2023), prediction of protein mutational effects (Ferruz and…”
BackgroundMethodInfluentialDART-Eval: A Comprehensive DNA Language Model Evaluation Benchmark on Regulatory DNA
Patel, A., Singhal, A., Wang, A. T., Pampari, A., Kasowski, M. M., & Kundaje, A. (2024) · Neural Information Processing Systems · doi:10.48550/arxiv.2412.05430 · Cited by 36
“In the biological domain, self-supervised language models trained on protein sequences have enabled high-performance tools for protein structure prediction and design [42, 29, 27].”
Protein Language Model Fitness Is a Matter of PreferencePreprint
Gordon, C., Lu, A. X., & Abbeel, P. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.10.03.616542 · Cited by 28
“In contrast, ProGen (Alley et al., 2019; Madani et al., 2023; Nijkamp et al., 2023) and other works embraced the GPT-like pretraining style and opted for autoregressive training scheme.”
ResultAI for organic and polymer synthesis
Hong, X., Yang, Q., Liao, K., Pei, J., Chen, M., Mo, F., Lu, H., Zhang, W.-B., Zhou, H., Chen, J., et al. (2024) · Science China Chemistry · doi:10.1007/s11426-024-2072-4 · Cited by 41
LLAMAFUZZ: Large Language Model Enhanced Greybox Fuzzing
Zhang, H., Rong, Y., He, Y., & Chen, H. (2026) · International Conference/Workshop on Automation of Software Test · doi:10.1145/3793654.3793756 · Cited by 30
The type V effectors for CRISPR/Cas-mediated genome engineering in plants
Zhang, R., Chai, N., Liu, T., Zheng, Z., Lin, Q., Xie, X., Wen, J., Yang, Z. F., Liu, Y., & Zhu, Q. (2024) · Biotechnology Advances · doi:10.1016/j.biotechadv.2024.108382 · Cited by 36
Liu, C., Yuan, Y., Yin, Y.-C., Xu, Y., Xu, X., Chen, Z., Wang, Y., Shang, L.-F., Liu, Q., & Zhang, M. (2025) · Annual Meeting of the Association for Computational Linguistics · doi:10.48550/arxiv.2506.04592 · Cited by 35
“The rapid advancement of LLMs has catalyzed transformative applications across diverse domains, ranging from medical diagnostics to fundamental scientific discovery (Jumper et al., 2021, 2020; Madani et al., 2023; Yang et al., 2024a; Feng et al., 2024; Shao et al., 2024; Ying et al., 2024b).”
Qian, H., Wang, Y., Zhou, X.-B., Gu, T., Wang, H., Lyu, H., Li, Z., Li, X., Zhou, H., Guo, C.-C., et al. (2025) · Nature Communications · doi:10.1038/s41467-025-58521-y · Cited by 35
“These results are consistent with fi ndings from other studies 14,37 and suggest that Euclidean distance could serve as a unique metric for assessing enzyme functionality, distinct from methods based on structural and sequence similarities alone.”
Engineering highly active nuclease enzymes with machine learning and high-throughput screening
Thomas, N., Belanger, D., Xu, C. A., Lee, H., Hirano, K., Iwai, K., Polic, V., Nyberg, K. D., Hoff, K. G., Frenz, L., et al. (2025) · Cell Systems · doi:10.1016/j.cels.2025.101236 · Cited by 41
Barman, K. G., Caron, S., Sullivan, E., de Regt, H. W., de Austri, R. R., Boon, M., Färber, M., Fröse, S., Golling, T., Lopez, L. G., et al. (2025) · The European Physical Journal C · doi:10.1140/epjc/s10052-025-14707-8 · Cited by 42
“…including the design and execution of autonomous experiments, agent-based AI labs, applications to symbolic tasks, and the development of collaborative AI tools known as CoScientists, and the enhancement of capabilities in programming and mathematical proof generation [22, 23, 24, 25, 26, 27, 28].”
Jin, R., Ye, Q., Wang, J., Cao, Z., Jiang, D., Wang, T., Kang, Y., Xu, W., Hsieh, C., & Hou, T. (2024) · Briefings in Bioinformatics · doi:10.1093/bib/bbae304 · Cited by 33
“Recently, there has been growing interest in the application of cutting-edge DL algorithms, such as Transformer-based BERT [28] and GPT [ 29], in diverse fields ranging from natural language processing to bioinformatics and computational biology [30–35].”
BackgroundZhang, Y., Liu, L., Xu, B., Zhang, Z., Yang, M. J., He, Y., Chen, J., Zhang, Y., Hu, Y., Chen, X., et al. (2024) · Acta Pharmaceutica Sinica B · doi:10.1016/j.apsb.2024.05.003 · Cited by 37
Self-citationSimultaneous enhancement of multiple functional properties using evolution-informed protein design
Fram, B., Su, Y., Truebridge, I., Riesselman, A. J., Ingraham, J. B., Passera, A., Napier, E., Thadani, N. N., Lim, S., Roberts, K., et al. (2024) · Nature Communications · doi:10.1038/s41467-024-49119-x · Cited by 34
Yang, E.-W., & Velazquez-Villarreal, E. I. (2025) · Bioinformatics · doi:10.1093/bioinformatics/btaf359 · Cited by 44
IgGM: A Generative Model for Functional Antibody and Nanobody DesignPreprint
Wang, R., Wu, F., Gao, X., Wu, J., Zhao, P., & Yao, J. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.09.19.613838 · Cited by 27
“In recent years, various machine learning methods have been applied to protein design, focusing on either sequences (Madani et al., 2023; Ferruz et al., 2022; Munsamy et al., 2022; Alamdari et al., 2023) or structures (Watson et al., 2023; Ingraham et al., 2023; Lutz et al., 2023).”
Zhou, B., Zheng, L., Wu, B., Yi, K., Zhong, B., Tan, Y., Liu, Q., Lió, P., & Hong, L. (2024) · Cell Discovery · doi:10.1038/s41421-024-00728-2 · Cited by 36
Generative models for protein structures and sequences
Hsu, C., Fannjiang, C., & Listgarten, J. (2024) · Nature Biotechnology · doi:10.1038/s41587-023-02115-w · Cited by 32
Open-endedness in synthetic biology: A route to continual innovation for biological design
Stock, M., & Gorochowski, T. E. (2024) · Science Advances · doi:10.1126/sciadv.adi3621 · Cited by 34
“Recent advances in deep learning–assisted protein design “hallucinate” a wide diversity of protein scaffolds to engineer such as new- to-nature luciferases ( 80 ) or enzybiotics ( 81 ).”
The role and pathway of VQ family in plant growth, immunity, and stress response
Tian, J., Zhang, J., & Francis, F. (2023) · Planta · doi:10.1007/s00425-023-04292-z · Cited by 34
“…applied in protein structure prediction and new functional protein development (Baek et al. 2021; Tunyasuvunakool et al. 2021; Bryant et al. 2022; Dauparas et al. 2022; Wang et al. 2022b; Huang et al. 2023; Lutz et al. 2023; Madani et al. 2023; Schuster et al. 2023; Watson et al. 2023).”
BackgroundDecoding the Foodome: Molecular Networks Connecting Diet and Health
Menichetti, G., Barabási, A.-L., & Loscalzo, J. (2024) · Annual Review of Nutrition · doi:10.1146/annurev-nutr-062322-030557 · Cited by 32
Bashour, H., Smorodina, E., Pariset, M., Zhong, J., Akbar, R., Chernigovskaya, M., Lê Quý, K., Snapkow, I., Rawat, P., Krawczyk, K., et al. (2024) · Communications Biology · doi:10.1038/s42003-024-06561-3 · Cited by 32
“(ii ) The second embedding is an amino acid sequence encoding output produced by the protein language model (PLM) ESM-1v, which generates large semantically-rich digital representations of antibodies (high-dimensional embedding — order of magnitude ≈ 10 3 ) 74 – 76 .”
PTM-Mamba: a PTM-aware protein language model with bidirectional gated Mamba blocks
Peng, F. Z., Wang, C.-T., Chen, T., Schussheim, B., Vincoff, S., & Chatterjee, P. (2025) · Nature Methods · doi:10.1038/s41592-025-02656-9 · Cited by 32
Who Determines What Is Relevant? Humans or AI? Why Not Both?
Faggioli, G., Dietz, L., Clarke, C. L. A., Demartini, G., Hagen, M., Hauff, C., Kando, N., Kanoulas, E., Potthast, M., Stein, B., et al. (2024) · Communications of the ACM · doi:10.1145/3624730 · Cited by 46
“In recent empirical studies on Web search 3 but also in programming, 7 human–computer interaction, 5 or protein function prediction, 10 it has been shown that LLM-generated opinions often agree with the assessment of humans.”
BackgroundProgrammable enzyme catalysis based on multiscale confinements
Cao, Y., Qiao, W., Zhang, C., Zhu, M., Wu, Q., Lou, W.-Y., Willner, I., Zare, R. N., & Ge, J. (2025) · Nature Synthesis · doi:10.1038/s44160-025-00898-1 · Cited by 34
Biomaterials in cellular agriculture and plant-based foods for the future
Gordon, E. B., Choi, I., Amanipour, A., Hu, Y., Nikkhah, A., Koysuren, B., Jones, C., Nitin, N., Ovissipour, R., Buehler, M. J., et al. (2025) · Nature Reviews Materials · doi:10.1038/s41578-025-00800-7 · Cited by 33
Leveraging Biomolecule and Natural Language through Multi-Modal Learning: A SurveyPreprint
Pei, Q., Zhou, Z., Gao, K., Jinhua, Z., Wang, Y., Wang, Z., Qin, T., Lijun, W., & Yan, R. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2403.01528 · Cited by 28
Accurate and robust protein sequence design with CarbonDesign
Ren, M., Yu, C., Bu, D., & Zhang, H. (2024) · Nature Machine Intelligence · doi:10.1038/s42256-024-00838-2 · Cited by 30
“For example, ProGen2 employs a generative pretrained transformer model to generate sequences with control tags specifying protein properties 50 .”
Programmable design of functional proteins from natural languagePreprint
Dai, F.-Y., You, S.-Y., Zhu, Y.-D., Gao, Y., Fu, L.-H., Zhou, X.-B., Su, J., Wang, C.-T., Fan, Y.-L., Tan, Q.-S., et al. (2026) · bioRxiv · doi:10.1101/2024.08.01.606258 · Cited by 30
“Recent advances in artificial intelligence (AI), particularly deep generative models, have enabled an emerging design paradigm in which models learn from protein sequences 14–16 , structures 17–24 , and functional annotations 25,26 to generate proteins under diverse design objectives.”
Artificial intelligence-driven computational methods for antibody design and optimization
Vecchietti, L. F., Wijaya, B. N., Armanuly, A., Hangeldiyev, B., Jung, H., Lee, S., Cha, M., & Kim, H. M. (2025) · mAbs · doi:10.1080/19420862.2025.2528902 · Cited by 32
“Prediction methods for antibody-related developability properties are further discussed in existing literature, 82,161,162 although the development of these prediction methods is largely hindered by limited data availability and the inconsistency of experimental protocols and measurements used to…”
Mantena, S., Pillai, P. P., Petros, B. A., Welch, N. L., Myhrvold, C., Sabeti, P. C., & Metsky, H. C. (2024) · Nature Biotechnology · doi:10.1038/s41587-024-02422-w · Cited by 40
“Generative methods are transforming the field of biological sequence design, with recent work focusing on generating long protein and DNA sequences 10,12,45 .”
DeepMineLys: Deep mining of phage lysins from human microbiome
Fu, Y., Yu, S., Li, J., Lao, Z., Yang, X., & Lin, Z. (2024) · Cell Reports · doi:10.1016/j.celrep.2024.114583 · Cited by 31
Structure-informed protein language models are robust predictors for variant effects
(2024) · doi:10.1007/s00439-024-02695-w · Cited by 30
“Since then, a series of transformer-based pLMs have been developed with increasing size and complexity, from 10 8 to 10 10 parameters, represented by ESM-1b (Rives et al. 2019), ProGen (Madani et al. 2023), ProtTrans (Elnaggar et al. 2021), Tranception (Notin et al. 2022), MSA-transformer (Rao et al. 2021), ESM-1v (Me…”
Artificial intelligence driven protein design and sustainable nanomedicine for advanced theranostics
Esmaeilpour, D., Hamblin, M. R., Cheng, J., Khosravi, A., Liu, J., Zarepour, A., Zarrabi, A., Sillanpaa, M., Zare, E. N., Shen, J., et al. (2026) · Bioactive Materials · doi:10.1016/j.bioactmat.2026.01.036 · Cited by 18
Generative AI: driving productivity and scientific breakthroughs in pharmaceutical R&D
Doron, G., Genway, S., Roberts, M., & Jasti, S. (2024) · Drug Discovery Today · doi:10.1016/j.drudis.2024.104272 · Cited by 39
Xu, K., Feng, H., Zhang, H., He, C., Kang, H., Yuan, T., Shi, L., Zhou, C., Hua, G., Cao, Y., et al. (2024) · Nature Biomedical Engineering · doi:10.1038/s41551-024-01220-8 · Cited by 36
Large-language models facilitate discovery of the molecular signatures regulating sleep and activity
Peng, D., Zheng, L., Liu, D., Han, C., Wang, X., Yang, Y. Q., Li, S., Zhao, M., Wei, Y., Li, J., et al. (2024) · Nature Communications · doi:10.1038/s41467-024-48005-w · Cited by 25
Diffusion on language model encodings for protein sequence generationPreprint
Meshchaninov, V., Страшнов, П. В., Shevtsov, A., Nikolaev, F., Ivanisenko, N. V., Kardymon, O. L., & Vetrov, D. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2403.03726 · Cited by 25
“Recent progress in this field has been driven by three main approaches: autoregressive models (Madani et al., 2023; Ferruz et al., 2022; Shin et al., 2021; Lv et al., 2024), which have demonstrated effectiveness in capturing sequential dependencies, diffusion models (Alamdari et al., 2023; Wang et…”
InfluentialFrom traditional to AI-driven: The evolution of intelligent enzyme engineering for biocatalysis
Lv, F., Zhang, J., You, S., & Qi, W. (2025) · Biotechnology Advances · doi:10.1016/j.biotechadv.2025.108788 · Cited by 31
Mixture of Experts Enable Efficient and Effective Protein Understanding and DesignPreprint
Sun, N., Zou, S., Tao, T., Mahbub, S., Li, D., Zhuang, Y., Wang, H., Cheng, X., Song, L., & Xing, E. P. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.11.29.625425 · Cited by 26
“The huge success of large language models (LLMs) in natural language processing (NLP) inspire researchers to apply self-supervised pre-training in the protein domain, using protein sequences without any structural and functional labels [5, 6, 7, 8, 9, 10, 11, 12, 13, 3, 14, 15, 4].”
AI-driven antibody design with generative diffusion models: current insights and future directions
He, X., Li, J.-R., Xu, J., Shan, H., Shen, S.-Y., Gao, S.-H., & Xu, H. E. (2024) · Acta Pharmacologica Sinica · doi:10.1038/s41401-024-01380-y · Cited by 28
“Similar to text generation tasks, language models have been prominently used in protein design by masking residues or more generally, tokens [22, 23].”
A review of ophthalmology education in the era of generative artificial intelligence
Heinke, A., Radgoudarzi, N., Huang, B. B., & Baxter, S. L. (2024) · Asia-Pacific Journal of Ophthalmology · doi:10.1016/j.apjo.2024.100089 · Cited by 39
Advances in stress-tolerance elements for microbial cell factories
Kuang, Z., Yan, X., Yuan, Y., Wang, R., Zhu, H., Wang, Y., Li, J., Ye, J., Yue, H., & Yang, X. (2024) · Synthetic and Systems Biotechnology · doi:10.1016/j.synbio.2024.06.008 · Cited by 37
“Additionally, universal LLM ( e.g. , ChatGPT4) also shows great potential in biology research [187], which is also used for effective recognition, prediction, and design of biological elements [188].”
BackgroundPosition: Leverage Foundational Models for Black-Box OptimizationPreprint
Song, X., Tian, Y., Lange, R. T., Lee, C., Tang, Y., & Chen, Y. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2405.03547 · Cited by 20
“For instance, techniques from protein design (Castro et al., 2022; Madani et al., 2023) train foundation models over protein sequence data as a form of knowledge absorption.”
MethodProtChatGPT: Towards Understanding Proteins with Large Language ModelsPreprint
Wang, C., Fan, H., Quan, R., & Yang, Y. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2402.09649 · Cited by 39
“Considering protein sequences as language in life, several works (Madani et al., 2023; Notin et al., 2022) encode amino acid tokens using the Transformer model (Vaswani et al., 2017) to extract pair-wise relationships among amino acids, and autoregressively recover protein sequences on extensive…”
MethodComputational Design of Peptide Assemblies
Min, J., Rong, X., Zhang, J., Su, R., Wang, Y., & Qi, W. D. (2024) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.3c01054 · Cited by 31
T-cell receptor binding prediction: A machine learning revolution
Weber, A., Pélissier, A., & Martínez, M. R. (2024) · ImmunoInformatics · doi:10.1016/j.immuno.2024.100040 · Cited by 28
“…the year 2020, a variety of Transformer models have emerged that, trained on extensive protein sequence datasets, are demonstrating competitive performance in various protein-related tasks, such as predicting protein homology, structure, function, and interactions [68, 69, 70, 71, 72, 73, 74].”
BackgroundProtein Language Models: Applications and Perspectives
Leclercq, M., & Droit, A. (2025) · Journal of Proteome Research · doi:10.1021/acs.jproteome.5c00506 · Cited by 29
Multimodal integration strategies for clinical application in oncology
Zhang, B., Wan, Z., Luo, Y., Zhao, X., Samayoa, J. A., Zhao, W., & Wu, S. (2025) · Frontiers in Pharmacology · doi:10.3389/fphar.2025.1609079 · Cited by 28
Large Language Models in Bioinformatics: A Survey
Wang, Z.-Y., Wang, Z.-K., Jiang, J.-Y., Chen, P.-A., Shi, X.-Y., & Li, Y. (2025) · Annual Meeting of the Association for Computational Linguistics · doi:10.48550/arxiv.2503.04490 · Cited by 27
Metadata
- Authors
- Madani, A., Krause, B., Greene, E. R., Subramanian, S., Mohr, B. P., Holton, J. M., Olmos, J. L., Xiong, C., Sun, Z. Z., Socher, R., Fraser, J. S., & Naik, N.
- Affiliations
- Salesforce (United States) (US)
- University of California, San Francisco (US)
- Howard Hughes Medical Institute (US)
- Lawrence Berkeley National Laboratory (US)
- Corresponding author
- Madani, Ali
- Naik, Nikhil
- ORCID
- Venue
- Nature Biotechnology
- Published
- 2023-01-26
- Volume
- 41
- Issue
- 8
- Pages
- 1099-1106
- ISSN
- 0733-222X (ISSN-L)
- 1087-0156
- 1546-1696
- PMC
- PMC10400306
- OpenAlex
- W4318071656
- Cited by
- 1,078
- References
- 80
- Citations by year
- 2026: 314 · 2025: 320 · 2024: 302 · 2023: 141
- FWCI
- 95.97
- Type
- journal article
- Language
- en
- Open access
- green
- www.ncbi.nlm.nih.gov
- Licence
- Locations
- Nature Biotechnology (publishedVersion)
- PubMed (publishedVersion)
- PubMed Central (acceptedVersion)
- eScholarship (California Digital Library) (submittedVersion)
- OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information) (submittedVersion)
- Funders
- U.S. Department of Health & Human Services | National Institutes of Health (GM123159)
- 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 › Molecular Biology › Enzyme Catalysis and Immobilization
- Keywords
- large language models, protein sequence generation, protein design, protein families, lysozyme, chorismate mutase, malate dehydrogenase, catalytic efficiency, controllable generation, protein engineering
- MeSH terms
- Amino Acid Sequence, Chorismate Mutase, Estrogens, Conjugated (USP), Language, Proteins
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 2026-10-05, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-05 (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
