The DOI matches its Crossref record; 77 of the 77 references it cites were found in a registry or an index; none of the 23 sources its claims cite is checked for support yet.
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 · 12 Not checked yet · 8 Own method · 4 No basis given
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.
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
- Not checked yetcites [14] Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021). source
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
“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 301–400 of 1,012
AlphaEvolve: A coding agent for scientific and algorithmic discovery
Novikov, A., Vũ, N., Eisenberger, M., Dupont, E., Huang, P.-S., Wagner, A. Z., Shirobokov, S., Kozlovskii, B., Ruiz, F. J. R., Mehrabian, A., et al. (2025) · arXiv · doi:10.48550/arxiv.2506.13131 · Cited by 15
“In particular, there are numerous recent LLM-based methods that target scientific problems in multiple disciplines, such as materials science [45, 71, 94, 119], chemistry [12, 64], bioinformatics [67, 85], geoscience [79], and quantum physics [30, 78] (for surveys on the topic, see [36, 65, 81]).”
Large Language Models for manufacturing
Li, Y., Zhao, H., Jiang, H., Pan, Y., Liu, Z., Wu, Z., Shu, P., Tian, J., Yang, T., Xu, S., et al. (2026) · Journal of Manufacturing Systems · doi:10.1016/j.jmsy.2026.02.014 · Cited by 108
Guiding Reasoning in Small Language Models with LLM AssistancePreprint
Kim, Y., Yi, E., Kim, M., Yun, S.-Y., & Kim, T. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2504.09923 · Cited by 20
“The significant success of large language models (LLMs) in natural language processing (NLP) has driven researchers and engineers to explore their potential applications in robotics (Driess et al., 2023) and fundamental science (Madani et al., 2023; Thirunavukarasu et al., 2023).”
Enhancing nucleotide sequence representations in genomic analysis with contrastive optimization
Refahi, M. S., Sokhansanj, B. A., Mell, J. C., Brown, J. R., Yoo, H., Hearne, G., & ROSEN, G. L. (2025) · Communications Biology · doi:10.1038/s42003-025-07902-6 · Cited by 14
ProtTeX: Structure-In-Context Reasoning and Editing of Proteins with Large Language Models
Ma, Z.-C., Fan, C., Wang, Z., Chen, Z.-Y., Lin, X.-H., Li, Y.-H., Feng, S., Cao, Z.-Q., Zhang, J., & Gao, Y.-Q. (2025) · Journal of Chemical Information and Modeling · doi:10.48550/arxiv.2503.08179 · Cited by 15
“9 The transformative impact of artificial intelligence on protein science was further underscored by the 2024 Nobel Prize in Chemistry, which recognized pioneering advancements in protein engineering.”
Zhu, S.-Y., Li, N., Liu, Z.-H., Yuan, Y.-J., & Li, B.-Z. (2025) · Green Carbon · doi:10.1016/j.greenca.2024.11.005 · Cited by 15
Twenty years of advances in prediction of nucleic acid-binding residues in protein sequences
Basu, S., Jing, Y., Kihara, D., & Kurgan, L. (2024) · Briefings Bioinform. · doi:10.1093/bib/bbaf016 · Cited by 15
Shah, A., Guntuboina, C., & Farimani, A. B. (2024) · arXiv.org · doi:10.48550/arxiv.2410.19222 · Cited by 15
“14 ProtGPT2 is a language model based on GPT 2 15 which has been trained on the protein space to generative novel proteins which follow the principle of natural ones.”
BackgroundLiao, M., Feng, S., Liu, X., Xu, G., Li, S., Bai, Y., Luo, H., Yao, B., Wang, H., & Tu, T. (2024) · Advanced Science · doi:10.1002/advs.202402441 · Cited by 12
Reconstruction of Ancestral Protein Sequences Using Autoregressive Generative Models
Leonardis, M. D., Pagnani, A., & Barrat-Charlaix, P. (2024) · bioRxiv · doi:10.1093/molbev/msaf070 · Cited by 15
Multi-Modal CLIP-Informed Protein Editing
Yin, M.-Z., Zhou, H., Zhu, Y., Lin, M., Wu, Y., Wu, J., Xu, H.-X., Hsieh, C.-Y., Hou, T.-J., Chen, J.-T., et al. (2024) · bioRxiv · doi:10.1101/2024.07.28.605469 · Cited by 15
“Recently, machine learning-based methods have demonstrated great promise in a wide range of protein-related applications, including 3D structure prediction [4, 5], mutation effects prediction [6], functionality prediction [7, 8, 9] and de novo protein design [10, 11].”
BackgroundMethodBeware of Data Leakage from Protein LLM PretrainingPreprint
Hermann, L., Fiedler, T., Nguyen, H., Nowicka, M., & Bartoszewicz, J. M. (2024) · bioRxiv · doi:10.1101/2024.07.23.604678 · Cited by 15
“Autoregressive PLMs are in turn capable of both conditional and unconditional protein sequence generation, including finetuning for target properties [21, 8, 24, 22, 13, 23].”
BackgroundDe novo design of buttressed loops for sculpting protein functions
Jiang, H., Jude, K. M., Wu, K., Fallas, J. A., Ueda, G., Brunette, T., Hicks, D. R., Pyles, H., Yang, A., Carter, L. P., et al. (2024) · Nature Chemical Biology · doi:10.1038/s41589-024-01632-2 · Cited by 24
“…alone or integrated with additional recent progress in loop design 16 and recently developed deep learning approaches for protein design 17–21 (which do not directly address the challenge of designing structured long loops), should enable the design of structured loops for binding…”
BackgroundOpen‐source large language models in action: A bioinformatics chatbot for PRIDE database
Bai, J., Kamatchinathan, S., Kundu, D. J., Bandla, C., Vizcaíno, J. A., & Pérez‐Riverol, Y. (2024) · PROTEOMICS · doi:10.1002/pmic.202400005 · Cited by 16
AI for Biomedicine in the Era of Large Language ModelsPreprint
Bi, Z., Dip, S. A., Hajialigol, D., Kommu, S., Liu, H., Meng, L., & Wang, X. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2403.15673 · Cited by 20
Sequence, Structure, and Functional Space of Drosophila De Novo Proteins
Middendorf, L., Iyengar, B. R., & Eicholt, L. A. (2024) · Genome Biology and Evolution · doi:10.1093/gbe/evae176 · Cited by 15
Wang, Y., Lv, H., Lei, R., Yeung, Y.-H., Shen, I. R., Choi, D., Teo, Q. W., Tan, T. J. C., Gopal, A. B., Chen, X., et al. (2024) · Biophysical Journal · doi:10.1016/j.bpj.2023.11.3334 · Cited by 8
When Protein Structure Embedding Meets Large Language Models
Ali, S., Chourasia, P., & Patterson, M. (2023) · Genes · doi:10.3390/genes15010025 · Cited by 15
“To predict a protein function, information from its structure or sequence can be extracted with the aid of language models [18,24,25].”
MethodA Data-Driven Approach for the Identification of Features for Automated Feedback on Academic Essays
Abbas, M., van Rosmalen, P., & Kalz, M. (2023) · IEEE Transactions on Learning Technologies · doi:10.1109/tlt.2023.3320877 · Cited by 16
Integrating MHC Class I visibility targets into the ProteinMPNN protein design process
Gasser, H.-C., Oyarzún, D. A., Alfaro, J., & Rajan, A. (2024) · doi:10.1101/2024.06.04.597365 · Cited by 15
“The current machine learning (ML) revolution is leading to a plethora of data-driven approaches using for example LLMs [14, 15], Variational Autoencoders (VAEs) [16] and Generative Adversarial Networks (GANs) [17] to generate new amino-acid sequences.”
MethodDeep learning guided design of protease substrates
Martin-Alonso, C., Alamdari, S., Samad, T. S., Yang, K. K., Bhatia, S. N., & Amini, A. P. (2026) · Nature Communications · doi:10.1038/s41467-025-67226-1 · Cited by 14
“LSTMs learn patterns sequentially, while trans-formers learn patterns by looking at all elements of a sequence simultaneously and are the dominant architecture for protein language modeling 52,53 .”
Unveiling the risks of ChatGPT in diagnostic surgical pathology
Guastafierro, V., Corbitt, D. N., Bressan, A., Fernandes, B., Mintemur, Ö., Magnoli, F., Ronchi, S., La Rosa, S., Uccella, S., & Renne, S. L. (2025) · Virchows Archiv · doi:10.1007/s00428-024-03918-1 · Cited by 14
Ibisanmi, T. A., Jiang, X., Willcox, M. D. P., & Kumar, N. (2026) · Frontiers in Bioinformatics · doi:10.3389/fbinf.2026.1749404 · Cited by 18
“Some recent papers address this by explicitly testing models on “out of distribution” sequences and by engineering novelty promoting regularization (Porto et al., 2017; Das et al., 2021; Madani et al., 2023).”
Multi-dimensional metagenomics
Peng, H., Ruiz-Moreno, A. J., & Fu, J. (2025) · Nature Reviews Bioengineering · doi:10.1038/s44222-025-00346-x · Cited by 19
Exploring the space of self-reproducing ribozymes using generative models
Lambert, C. N., Opuu, V., Calvanese, F., Pavlinova, P., Zamponi, F., Hayden, E. J., Weigt, M., Smerlak, M., & Nghe, P. (2025) · Nature Communications · doi:10.1038/s41467-025-63151-5 · Cited by 12
“In the long term, generating RNAs with the same activity but much less sequence identity or without length constraints should be possible, and has been demonstrated in proteins 49 .”
When Efficiency Meets Fulfillment: Understanding Long-Term LLM Integration in Knowledge Work
Kobiella, C., Mitrevska, T., Schmidt, A., & Draxler, F. (2025) · Symposium on Human-Computer Interaction for Work · doi:10.1145/3729176.3729180 · Cited by 14
Co-design protein sequence and structure in discrete space via generative flow
Yang, S., Ju, L., Cheng, P., Zhou, J., Cai, Y., & Feng, D. (2025) · Bioinformatics · doi:10.1093/bioinformatics/btaf248 · Cited by 14
“…(Ho et al. 2020), have been introduced for protein design, leading to significant advancements in both protein sequence design (Dauparas et al. 2022, Ferruz et al. 2022, Madani et al. 2023) and structure design (Watson et al. 2023, Yim et al. 2023, Wang et al. 2024) (Supplementary Section S1).”
Discovery of diverse and high-quality mRNA capping enzymes through a language model–enabled platform
Wang, T., Qin, B. R., Li, S. J., Wang, Z., Li, X., Jiang, Y., Qin, C., Ouyang, Q., Lou, C., & Qian, L. (2025) · Science Advances · doi:10.1126/sciadv.adt0402 · Cited by 9
“…useful synbio-parts similar to CEs and to unveil the natural diversity hidden in the vast literature base for subsequent learning and design ( 21 , 22 ), it is important to establish an automatic mining pipeline that exploits both literature and databases for evidence-based gene discovery (note…”
Self‐Driving Microscopes: AI Meets Super‐Resolution Microscopy
Ward, E. N., Scheeder, A., Barysevich, M., & Kaminski, C. (2025) · Small Methods · doi:10.1002/smtd.202401757 · Cited by 14
Protein Engineering of Substrate Specificity toward Nitrilases: Strategies and Challenges.
Bian, S.-Q., Wang, Z.-K., Gong, J.-S., Su, C., Li, H., Xu, Z.-H., & Shi, J.-S. (2025) · Journal of Agricultural and Food Chemistry · doi:10.1021/acs.jafc.4c09599 · Cited by 14
Integrating genetic algorithms and language models for enhanced enzyme design
Teukam, Y. G. N., Zipoli, F., Laino, T., Criscuolo, E., Grisoni, F., & Manica, M. (2024) · Briefings in Bioinformatics · doi:10.1093/bib/bbae675 · Cited by 13
“By leveraging protein sequence databases, LLMs offer unprecedented insights into protein function and interactions, revolutionizing our understanding of biological systems [11].”
Flöge, K., Udayakumar, S., Sommer, J., Piraud, M., Kesselheim, S., Fortuin, V., Günnemann, S., van der Weg, K. J., Gohlke, H., Merdivan, E., et al. (2025) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1013679 · Cited by 13
“These enriched representations support a range of downstream tasks, including protein function prediction [45] and enzyme or antibody design [31], thereby advancing functional protein design [30].”
High-Temperature Tolerance Protein Engineering through Deep Evolution
Chu, H., Tian, Z., Hu, L., Zhang, H., Chang, H., Bai, J., Liu, D., Lu, L., Cheng, J., & Jiang, H. (2024) · BioDesign Research · doi:10.34133/bdr.0031 · Cited by 14
“Studies on protein sequence design have demonstrated that deep learning models can learn the diversity of natural protein sequences, enabling the generation of protein variants to expand the functional protein space [21,22,29,30].”
BackgroundMethodAll-Atom Biomolecular Simulation in the Exascale Era
Beck, T. L., Carloni, P., & Asthagiri, D. N. (2024) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.3c01276 · Cited by 15
Yang, X., Wuchty, S., Liang, Z., Ji, L., Wang, B., Zhu, J., Zhang, Z., & Dong, Y. (2024) · Briefings in Bioinformatics · doi:10.1093/bib/bbae005 · Cited by 16
“Moreover, large protein language models such as ESM and ProGen have been applied in various protein bioinformatics prediction tasks [56–58].”
MethodFine-tuning protein language models unlocks the potential of underrepresented viral proteomes
Sawhney, R., Ferrell, B. D., Dejean, T., Schreiber, Z., Harrigan, W., Polson, S. W., Wommack, K. E., & Belcaid, M. (2025) · PeerJ · doi:10.7717/peerj.19919 · Cited by 14
Fine-Tuning Protein Language Models Unlocks the Potential of Underrepresented Viral Proteomes
Sawhney, R., Ferrell, B., Dejean, T., Schreiber, Z., Harrigan, W., Polson, S. W., Wommack, K. E., & Belcaid, M. (2025) · doi:10.1101/2025.04.17.649224 · Cited by 14
“Following the approach described in Devlin et. al (2019), we masked 15% of tokens in each protein sequence at random, with masking probability of 80% for ESM2 and 100% for ProtT5 and ProGen2.”
InfluentialSingle-round evolution of RNA aptamers with GRAPE-LM
Zhang, J., Ju, Z., Tang, S., Liu, C., Cai, Y., Zeng, H., Meng, X., Liu, B., Zhang, Y., & Wang, Y. (2026) · Nature Biotechnology · doi:10.1038/s41587-026-03007-5 · Cited by 12
Zhang, S., Yu, L., Brekelmans, R., Tang, L., Asif, S., & Steeg, G. V. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2602.00286 · Cited by 14
“Autoregressive (AR) models have established themselves as the dominant paradigm for discrete sequence generation, powering breakthroughs in natural language processing (Achiam et al., 2023; Touvron et al., 2023) and biological sequence design (Madani et al., 2023).”
A survey on large language models in biology and chemistry
Ashyrmamatov, I., Gwak, S. J., Jin, S.-Y., Jun, I., Ucak, U. V., Lee, J. Y., & Lee, J. (2025) · Experimental & Molecular Medicine · doi:10.1038/s12276-025-01583-1 · Cited by 15
“ProGen 17 enables controlled protein sequence generation by incorporating conditioning tags into an autoregressive transformer architecture.”
Sampling Protein Language Models for Functional Protein DesignPreprint
Darmawan, J. T., Gal, Y., & Notin, P. (2025) · bioRxiv · doi:10.1101/2025.09.14.676087 · Cited by 13
“While the field has witnessed the emergence of several PLMs for protein design [1, 10, 13, 26, 23], efficient sampling methods have been understudied to date.”
Best Practices for Machine Learning-Assisted Protein Engineering
Herrera‐Rocha, F., Medina-Ortiz, D., Mauz, F., Pleiss, J., & Davari, M. D. (2025) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.5c01983 · Cited by 12
Artificial intelligence in de novo protein design
Yao, J., & Wang, X. (2025) · Medicine in Novel Technology and Devices · doi:10.1016/j.medntd.2025.100366 · Cited by 14
Bielska, W., Jaszczyszyn, I., Dudzic, P., Janusz, B., Chomicz, D., Wrobel, S., Greiff, V., Feehan, R., Adolf‐Bryfogle, J., & Krawczyk, K. (2025) · Frontiers in Immunology · doi:10.3389/fimmu.2025.1571371 · Cited by 17
Effective Gene Expression Prediction and Optimization from Protein Sequences
Liu, T., Zhang, Y., Li, Y., Xu, G., Gao, H., Wang, P., Tu, T., Luo, H., Wu, N., Yao, B., et al. (2025) · Advanced Science · doi:10.1002/advs.202407664 · Cited by 9
Luo, J., Zhao, K., Chen, J., Yang, C., Qu, F., Liu, Y., Jin, X., Yan, K., Zhang, Y., & Liu, B. (2024) · Genomics Proteomics & Bioinformatics · doi:10.1093/gpbjnl/qzae084 · Cited by 21
“The pLMs pre-trained on millions of protein sequences [41] can capture long-range dependencies of amino acid residues [42–44].”
Yin, M., Zhou, H., Wu, J., Zhu, Y., Zhan, Y., Kong, Z., Xu, H., Hsieh, C., Chen, J., Hou, T., et al. (2025) · Research · doi:10.34133/research.0721 · Cited by 12
“While the interpretation of attention mechanisms remains an active area of research, this analysis provides an intuitive way to gain profound comprehension of the model’s functional mechanism.”
Fourier Head: Helping Large Language Models Learn Complex Probability Distributions
Gillman, N., Aggarwal, D., Freeman, M., Singh, S., & Sun, C. (2024) · International Conference on Learning Representations · doi:10.48550/arxiv.2410.22269 · Cited by 13
“For example, they have been used in protein synthesis (Madani et al., 2023), time series forecasting (Ansari et al., 2024; Das et al., 2024; Jin et al., 2024; Nate Gruver & Wilson, 2023; Requeima et al., 2024; Jia et al., 2024; Zhou et al., 2023; Wang et al., 2024), music generation (Dhariwal et…”
Cutting-edge computational approaches in enzyme design and activity enhancement
Sun, R., Wu, D., Chen, P., & Zheng, P. (2024) · Biochemical Engineering Journal · doi:10.1016/j.bej.2024.109510 · Cited by 20
Advancing plant biology through deep learning-powered natural language processing
Peng, S., & Rajjou, L. (2024) · Plant Cell Reports · doi:10.1007/s00299-024-03294-9 · Cited by 14
“Kotsiliti (2022) and Madani et al. (2023) illustrate PLMs’ role in advancing protein structure and function prediction, thereby enhancing our understanding of protein folding dynamics and biological functionalities.”
Mixture-of-Modules: Reinventing Transformers as Dynamic Assemblies of Modules
Gong, Z.-C., Lv, A., Guan, J., Wu, W., Zhang, H.-S., Huang, M., Zhao, D., & Yan, R. (2024) · Conference on Empirical Methods in Natural Language Processing · doi:10.48550/arxiv.2407.06677 · Cited by 13
“In addition, the success of Transformer in natural language processing (NLP) is also inspiring innovations in other fields such as computer vision (Peebles and Xie, 2023; Agostinelli et al., 2023) and biomedicine (Singhal et al., 2023; Madani et al., 2023).”
BackgroundMultifacetEval: Multifaceted Evaluation to Probe LLMs in Mastering Medical Knowledge
Zhou, Y., Liu, X., Ning, C., & Wu, J. (2024) · International Joint Conference on Artificial Intelligence · doi:10.48550/arxiv.2406.02919 · Cited by 13
“The rapid advancement of large language model (LLM) technology has achieved great success in various domains [Romera-Paredes et al. , 2023; Madani et al. , 2023; Boiko et al. , 2023].”
MethodZhang, P., Ma, J., & Chen, T. (2025) · IEEE Transactions on Computational Biology and Bioinformatics · doi:10.1109/tcbbio.2025.3576488 · Cited by 13
Compressing the collective knowledge of ESM into a single protein language model
Dinh, T., Jang, S.-K., Zaitlen, N., & Ntranos, V. (2026) · Nature Methods · doi:10.1038/s41592-026-03050-9 · Cited by 7
Generalist biological artificial intelligence in modeling the language of life
Rao, V. M., Zhang, S., Plosky, B. S., Hsu, P. D., Wang, B., Zou, J., Žitnik, M., Topol, E. J., & Rajpurkar, P. (2026) · Nature Biotechnology · doi:10.1038/s41587-026-03064-w · Cited by 14
Artificial intelligence‐driven anticancer peptide discovery
Wu, J.-R., Ji, S.-Q., Sahibzada, K., Lou, M., An, F.-Y., Li, W., Guo, J., Zhang, T., Zhang, X., Chou, Y., et al. (2025) · iMetaOmics · doi:10.1002/imo2.70063 · Cited by 12
“The embedding vectors generated by these models include the features of individual amino acids and reflect their positions and functional relevance within the sequence, offering richer representations [30].”
Self-citationPugh, C. W. J., Nuñez-Valencia, P. G., Dias, M., & Frazer, J. (2025) · bioRxiv · doi:10.1101/2025.05.20.655154 · Cited by 12
“…accuracy [Jumper et al., 2021, Baek et al., 2021], evaluating the effects of mutations [Meier et al., 2021], identifying functionally related proteins [Rives et al., 2021], and generating novel sequences with functional potential [Ferruz et al., 2022, Madani et al., 2023, Winnifrith et al., 2024].”
T cell receptor specificity landscape revealed through de novo peptide design
Visani, G. M., Pun, M. N., Minervina, A. A., Bradley, P. H., Thomas, P. G., & Nourmohammad, A. (2025) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2504783122 · Cited by 17
Self-Speculative Masked Diffusions
Campbell, A., Bortoli, V. D., Shi, J.-X., & Doucet, A. (2025) · arXiv.org · doi:10.48550/arxiv.2510.03929 · Cited by 12
“Generative models for discrete data are at the core of a wide range of modern deep learning systems from chatbots (Gemini Team, 2023; OpenAI Team, 2023; Anthropic, 2024) to biological foundation models (Hayes et al., 2025; Wang et al., 2025; Madani et al., 2023).”
Artificial intelligence for nanomedicine
Song, X., Gao, X., Wang, H., Yu, F., Qin, M., Li, Y., Liu, Y., Feng, W., Zhou, C., Chukavin, N. N., et al. (2025) · Science China Chemistry · doi:10.1007/s11426-025-2942-5 · Cited by 16
SafeProtein: Red-Teaming Framework and Benchmark for Protein Foundation Models
Fan, J.-G., Zhou, Z., Jin, R.-F., Cong, L., Wang, M.-D., & Zhang, Z.-X. (2025) · arXiv.org · doi:10.48550/arxiv.2509.03487 · Cited by 12
“[26], adopted autoregressive language modeling to de novo generate protein sequences.”
Wang, S.-Y., Bergman, M. T., Hall, C. K., & You, F.-Q. (2025) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.5c01401 · Cited by 12
AI4Protein: transforming the future of protein design
Wang, D., Tan, Z., Gao, J., Zhang, S., Shen, J., & Lu, Y. (2025) · Science China Life Sciences · doi:10.1007/s11427-024-2906-3 · Cited by 12
“Moreover, due to the scalability of transformer models, many protein foundational models (Lv et al., 2024; Madani et al., 2023) are built on this architecture, benefiting from its ability to process large datasets efficiently.”
PDFBench: A Benchmark for De novo Protein Design from Function
Kuang, J.-H., Liu, N.-W., Wang, J., Sun, C., Ji, T., & Wu, Y.-B. (2025) · arXiv.org · doi:10.48550/arxiv.2505.20346 · Cited by 12
“Unconditional de novo Unconditional design methods generate amino acid sequences without constraints, employing either autoregressive models (Madani et al., 2023; Ferruz et al., 2022; Hesslow et al., 2022) or discrete diffusion models (Alamdari et al., 2023; Wang et al., 2024b).”
Lin, L.-K., Xiao, D., Song, W., & Lu, W.-H. (2025) · Food Chemistry · doi:10.1016/j.foodchem.2025.143311 · Cited by 12
Applying Spatiotemporal Modeling of Cell Dynamics to Accelerate Drug Development
Chen, X., Xu, S., Chu, B., Guo, J., Zhang, H.-K., Sun, S.-Y., Song, L., & Feng, X. (2024) · ACS Nano · doi:10.1021/acsnano.4c12599 · Cited by 16
Estimating wage disparities using foundation models
Vafa, K., Athey, S., & Blei, D. (2024) · Proceedings of the National Academy of Sciences of the United States of America · doi:10.48550/arxiv.2409.09894 · Cited by 12
“…intractable prediction problems in domains such as computer vision (Dosovitskiy et al., 2021), music (Huang et al., 2019), and protein generation (Madani et al., 2023) A foundation model of labor market history can help estimate the unexplained wage gap by providing a low-dimensional…”
Applications of cell free protein synthesis in protein design
Thornton, E. L., Paterson, S. M., Stam, M. J., Wood, C. W., Laohakunakorn, N., & Regan, L. (2024) · Protein Science · doi:10.1002/pro.5148 · Cited by 15
“Madani and colleagues used large-language models to generate thousands of designs for lysozyme from five different structural classes (Madani et al., 2023).”
MethodPseudo-perplexity in One Fell Swoop for Protein Fitness Estimation
Kantroo, P., Wagner, G. P., & Machta, B. (2024) · arXiv.org · doi:10.48550/arxiv.2407.07265 · Cited by 12
“Normalizing the log-likelihood by the sequence length enables fitness comparisons across sequences with different lengths, elucidating why the heuristic also works for indel mutations [24].”
BackgroundPourmirzaei, M., Esmaili, F., Pourmirzaei, M., Wang, D.-L., & Xu, D. (2024) · bioRxiv · doi:10.1101/2024.05.31.596915 · Cited by 12
Large Language Model-Driven Urban Traffic Signal Control
Tang, Y., Dai, X., Zhao, C., Cheng, Q., & Lv, Y. (2024) · Australian & New Zealand Control Conference (ANZCC) · doi:10.1109/anzcc59813.2024.10432823 · Cited by 18
“…017000, China. tangyiqing2022@ia.ac.cn , xingyuan.dai@ia.ac.cn , zhaochen2020@ia.ac.cn , SQT2200402038@student.cumtb.edu.cn , yisheng.lv@ia.ac.cn are gradually participate in robot control [5], medical treat-ment [6], predicting protein sequences [7], and autonomous driving [8], [9], [10], [11].”
Background(2024) · doi:10.1101/2024.09.04.611278 · Cited by 12
“…(i.e. BERT-style self-supervised learning) or the next token in a sentence fragment (i.e. GPT-style self-supervised learning), protein language models are similarly trained to predict masked amino acid tokens (e.g., ProtBERT [12] and ESM [13]) or the next amino acid token (e.g., ProGen [14]).”
Computation and deep-learning-driven advances in CRISPR genome editing
Pindi, C., & Palermo, G. (2026) · Nature Structural & Molecular Biology · doi:10.1038/s41594-025-01739-7 · Cited by 10
Advances in the Modification of Enzymatic Properties Based on Protein Engineering Strategies
Huang, Y., Wen, L., & Yang, B. K. (2026) · Journal of Agricultural and Food Chemistry · doi:10.1021/acs.jafc.5c13434 · Cited by 11
Steering generative models for protein design: Aligning and conditioning strategies
Stocco, F., Garibbo, M., & Ferruz, N. (2026) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2026.103250 · Cited by 10
“SFT has achieved notable success, including the generation of enzymes [27, 28], gene editors[29, 30], or bacteriophages [31].”
Evaluating Pretrained Protein Language Model Embeddings as Proxies for Functional Similarity
Shaw, R., Love, S. D., & McWhite, C. D. (2025) · Journal of Molecular Evolution · doi:10.1007/s00239-025-10282-4 · Cited by 12
Enhancing functional proteins through multimodal inverse folding with ABACUS-T
Liu, Y., Wu, R., Wang, X., Wang, S., Chen, L.-H., Li, F., Chen, Q., & Liu, H. (2025) · Nature Communications · doi:10.1038/s41467-025-65175-3 · Cited by 9
“So far, successful adventures along this direction have relied on the availability of a very large number of diverse natural protein sequences sharing a common target function (the numbers of homologous sequences used ranged from 4,765 for malate dehydrogenases 12 to55,948forlysozymes 14 .”
Steering Protein Language Models
Huang, L.-K., Zhu, R.-Y., He, B., & Yao, J. (2025) · International Conference on Machine Learning · doi:10.48550/arxiv.2509.07983 · Cited by 11
“Recent advances leverage generative PLMs for de novo protein design [21, 25] and latent space optimization [35, 13].”
CFP-Gen: Combinatorial Functional Protein Generation via Diffusion Language Models
Yin, J.-B., Zha, C., He, W.-J., Xu, C.-C., & Gao, X. (2025) · International Conference on Machine Learning · doi:10.48550/arxiv.2505.22869 · Cited by 11
“The growing capabilities of large-scale protein language models (PLMs) offer tremendous potential for discovering novel proteins that do not exist in nature or would require millions of years of evolution to emerge (Hayes et al., 2025a; Madani et al., 2023; Alamdari et al., 2023).”
InfluentialEnhancing privacy in biosecurity with watermarked protein design
Chen, Y., Hu, Z., Wu, Y., Chen, R., Jin, Y., Zhan, M., Xie, C., Chen, W., & Huang, H. (2025) · Bioinformatics · doi:10.1093/bioinformatics/btaf141 · Cited by 9
“Among the computational methods, the capability of generative deep-learning approaches has rapidly grown and can perform various protein design tasks within a single model (Dauparas et al. 2022, Madani et al. 2023).”
Allman, B. E., Vieira, L., Diaz, D. J., & Wilke, C. O. (2025) · Journal of The Royal Society Interface · doi:10.1098/rsif.2024.0598 · Cited by 11
Biological databases in the age of generative artificial intelligence
Pop, M., Attwood, T. K., Blake, J., Bourne, P., Conesa, A., Gaasterland, T., Hunter, L. E., Kingsford, C., Kohlbacher, O., Lengauer, T., et al. (2025) · Bioinformatics Advances · doi:10.1093/bioadv/vbaf044 · Cited by 11
“AI technologies are also used in synthetic biology to generate new chemical compounds (Meier et al. 2020) or proteins (Ferruz et al. 2022, Madani et al. 2023).”
PointNSP: Autoregressive 3D Point Cloud Generation with Next-Scale Level-of-Detail Prediction
Meng, Z., Wang, Q.-C., Dou, Z., Song, Z., Zhou, Z.-P., King, I., & Zhao, P.-L. (2025) · Computer Vision and Pattern Recognition · doi:10.48550/arxiv.2510.05613 · Cited by 11
“Moreover, increasing the number of scales naturally leads to longer sampling times.”
Leveraging protein language models for robust antimicrobial peptide detection
Zhang, L., Xiong, S., Xu, L., Liang, J., Zhao, X., Zhang, H., & Tan, X. (2025) · Methods · doi:10.1016/j.ymeth.2025.03.002 · Cited by 12
Cell-free protein synthesis platforms for accelerating drug discovery
Caschera, F. (2025) · Biotechnology Notes · doi:10.1016/j.biotno.2025.02.001 · Cited by 11
“72 The model is a 1.2 billion neural network based on Transformer architecture with a self-attention mechanism that weights the probability of each token (amino acid).”
InfluentialXie, H., Liu, K., Li, Z., Wang, Z., Wang, C., Li, F., Han, W., & Wang, L. (2024) · JACS Au · doi:10.1021/jacsau.4c01045 · Cited by 14
Dharuman, G., Hippe, K., Brace, A., Foreman, S., Hatanpää, V., Sastry, V. K., Zheng, H., Ward, L. T., Muralidharan, S., Vasan, A. K., et al. (2024) · International Conference for High Performance Computing, Networking, Storage and Analysis (SC) · doi:10.1109/sc41406.2024.00013 · Cited by 9
“Unification of disparate knowledge (e.g., knowledge bases, sequence annotations) utilizing generative sequence models exists in very select task settings such as protein engineering for functional domains [25], [8] and text-guided protein engineering [10], [38], [39].”
InfluentialCOEFF-KANs: A Paradigm to Address the Electrolyte Field with KANs
Li, X.-H., Feng, Z., Chen, Y., Dai, W.-C., He, Z., Zhou, Y., & Jiao, S.-H. (2024) · arXiv.org · doi:10.48550/arxiv.2407.20265 · Cited by 11
“…Bajorath 2023), their practical implementation on further downstream tasks is still in its early stages and largely limited to sequencing problems such as predicting protein sequences (Madani et al. 2023), polymers (Chen, Tao, and Li 2021), and chemical reactions (Rankovi ´ c and Schwaller 2023).”
BackgroundImprovements in viral gene annotation using large language models and soft alignments
Harrigan, W., Ferrell, B. D., Wommack, K. E., Polson, S. W., Schreiber, Z. D., & Belcaid, M. (2024) · BMC Bioinformatics · doi:10.1186/s12859-024-05779-6 · Cited by 9
Pan, F., Liu, D., Tuersuntuoheti, T., Xing, H.-D., Zhu, Z., Fang, Y., Zhao, L., Zhao, L., Li, X.-X., Le, Y., et al. (2024) · Food Science of Animal Products · doi:10.26599/fsap.2024.9240053 · Cited by 11
In vivo functional phenotypes from a computational epistatic model of evolution
Alvarez, S., Nartey, C. M., Mercado, N., de la Paz, J. A., Huseinbegovic, T., & Morcos, F. (2024) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2308895121 · Cited by 20
PRO-LDM: Protein Sequence Generation with a Conditional Latent Diffusion ModelPreprint
Zhang, S., Jiang, Z., Huang, R., Mo, S., Zhu, L.-T., Li, P., Zhang, Z., Pan, E., Chen, X., Long, Y.-F., et al. (2024) · bioRxiv · doi:10.1101/2023.08.22.554145 · Cited by 11
De novo design of functional nucleic acids of aptamers
Zhang, Z., Jiang, M., He, A., Zhu, Y., Wang, E., Wan, L., Qiu, J., Guo, P., Chen, G., & Han, D. (2026) · Nature Computational Science · doi:10.1038/s43588-026-00965-3 · Cited by 14
“Using Large language models (LLMs) trained on a vast number of biological sequences have revolutionized the paradigm of protein design 19,25–27 , yet their application in nucleic acid design remains largely uncharted.”
The evolving landscape of pharmacogenomics: Current achievements and future directions
Lauschke, V. M., & Ingelman‐Sundberg, M. (2026) · Pharmacological Reviews · doi:10.1016/j.pharmr.2026.100122 · Cited by 13
Generative AI for synthetic biology: Designing biological parts, circuits, and genomes
Kim, N., Carluccio, G. D., Zhang, K., & Collins, J. J. (2026) · Cell Systems · doi:10.1016/j.cels.2026.101533 · Cited by 5
Sequence-based generative AI design of versatile tryptophan synthases
Lambert, T., Tavakoli, A., Dharuman, G., Yang, J., Bhethanabotla, V. C., Kaur, S., Hill, M., Ramanathan, A., Anandkumar, A., & Arnold, F. H. (2026) · Nature Communications · doi:10.1038/s41467-026-68384-6 · Cited by 11
“Recent experimental applications have validated this approach: PLMs can generate functional proteins with real-world relevance, marking a turning point for machine learning in protein engineering 13,19 – 22 .”
Protein foundation models: a comprehensive survey
Xu, H., Li, L., Pan, S., Cheng, P., Wang, Y., Rong, Z., Feng, L., Huang, X., Wang, S., & Shu, W. (2026) · Science China Life Sciences · doi:10.1007/s11427-025-3147-2 · Cited by 8
TCRdesign: an antigen-specific generative language model for de novo design of T-cell receptors
Li, X., Yang, Q., Xu, L., Dong, W., Wang, K., Dong, S., Wang, W., Luo, G., Zhang, X., Yang, T., et al. (2025) · Briefings in Bioinformatics · doi:10.1093/bib/bbaf691 · Cited by 14
“Models such as ESM [19], ProtTrans [20], and ProGen [21], which are pretrained on billions of natural sequences, capture evolutionary constraints, translating into accurate predictions of structures, sequences, functionality, and mutational fitness [22].”
Breaking Evolution’s Ceiling: AI-Powered Protein Engineering
Jin, S., Wu, Q., Fu, G., Lu, D. X., Wang, F., Deng, L. F., & Nie, K. (2025) · Catalysts · doi:10.3390/catal15090842 · Cited by 11
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).
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
