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 501–600 of 1,012
Nguyen, V. T. D., Nguyen, N. D., & Hy, T. S. (2024) · Structural Dynamics · doi:10.1063/4.0000271 · Cited by 8
ProteinAligner: A Multi-modal Pretraining Framework for Protein Foundation ModelsPreprint
Zhang, L., Guo, H., Schaffer, L. V., Ko, Y. S., Singh, D., Rahmani, H., Grotjahn, D., Villa, E., Gilson, M., Wang, W., et al. (2024) · bioRxiv · doi:10.1101/2024.10.06.616870 · Cited by 6
ConoDL: a deep learning framework for rapid generation and prediction of conotoxins
Guo, M., Li, Z., Deng, X., Luo, D., Yang, J., Chen, Y., & Xue, W. (2024) · Journal of Computer-Aided Molecular Design · doi:10.1007/s10822-024-00582-0 · Cited by 7
From predicting to decision making: Reinforcement learning in biomedicine
Liu, X., Zhang, J., Hou, Z., Yang, Y., & Gao, Y. Q. (2024) · Wiley Interdisciplinary Reviews Computational Molecular Science · doi:10.1002/wcms.1723 · Cited by 9
“With training on the scientific data, LLM has great potential to be constructed as general AI model in the field of chemistry, 124 biology 125,126 and clinics.”
Enhancing Biosecurity with Watermarked Protein DesignPreprint
Chen, Y.-S., Hu, Z.-M., Wu, Y.-H., Chen, R.-B., Jin, Y., Chen, W., & Huang, H. (2024) · bioRxiv · doi:10.1101/2024.05.02.591928 · Cited by 6
“Among the computational meth-ods, the capability of generative deep-learning approaches has rapidly grown and can perform various protein design tasks within a single model [2, 3].”
BackgroundAMPCliff: quantitative definition and benchmarking of activity cliffs in antimicrobial peptides
Li, K.-W., Wu, Y., Li, Y., Guo, Y., Wang, Y., Liang, Y.-Y., Fan, Y., Huang, L., Zhang, R.-C., & Zhou, F.-F. (2024) · arXiv.org · doi:10.48550/arxiv.2404.09738 · Cited by 6
“Sequence identity is frequently used in protein design to demonstrate the novelty of discovered proteins 41,61 , and can be measured using MMseqs2 62 , a widely-used clustering method for large protein datasets such as UniProt 63 .”
MethodComputational peptide discovery with a genetic programming approach
Scalzitti, N., Miralavy, I., Korenchan, D. E., Farrar, C. T., Gilad, A. A., & Banzhaf, W. (2024) · Journal of Computer-Aided Molecular Design · doi:10.1007/s10822-024-00558-0 · Cited by 9
“Another example is ProGen [92], a language models used to generate sequences of artificial proteins.”
Large language models for biomolecular analysis: From methods to applications
Feng, R., Zhang, C., & Zhang, Y. (2024) · TrAC Trends in Analytical Chemistry · doi:10.1016/j.trac.2024.117540 · Cited by 7
Intelligent Protein Design and Molecular Characterization Techniques: A Comprehensive Review
Wang, J., Chen, C., Yao, G., Ding, J., Wang, L., & Jiang, H. (2023) · Molecules · doi:10.3390/molecules28237865 · Cited by 10
“…successors GPT-2 and GPT-3 [170–172], the Evolutionary Scale Modeling (ESM) family for predicting protein structure and function (ESM-1b, ESM-MSA-1b, & ESM-1v) [173–175], the ProtTrans with the largest training dataset [176], and the ProGen language model [177] that can control protein generation.”
BackgroundMiniaturizing and modifying natural proteins with Raygun
Devkota, K., Shonai, D., Mao, Y., Ko, Y. S., Wang, W., Soderling, S. H., & Singh, R. (2026) · Nature · doi:10.1038/s41586-026-10842-8 · Cited by 12
“Protein language models (PLMs) encode proteins as rich, per-residue representations learned from evolutionary data 1,2,9–11 .”
Decoding complexity through machine learning is redefining scientific discovery
Vinuesa, R., Cinnella, P., Rabault, J., Azizpour, H., Bauer, S., Brunton, B. W., Elofsson, A., Elias Jarlebring, A., Kjellström, H., Markidis, S., et al. (2026) · Communications Physics · doi:10.1038/s42005-026-02676-7 · Cited by 17
“AlphaFoldhasalsospurredinnovationsinML,including single-sequencemethodslikeESMfoldandgenerativemodelsforprotein design 106,107 .”
Expanding functional protein sequence space using high entropy generative models
Netti, R., Hinds, E. C., Calvanese, F., Ranganathan, R., Weigt, M., & Zamponi, F. (2026) · arXiv.org · doi:10.48550/arxiv.2605.03578 · Cited by 5
“These architectures, often trained on massive, unaligned, and multi-family datasets, represent a paradigm shift in generative modeling [4–6, 44”
On using large language models to support social research for climate action
Nabavi, E., Maier, H. R., Razavi, S., Hindes, A., Howden, M. S., Grant, W. J., & Raman, S. (2026) · npj Climate Action · doi:10.1038/s44168-026-00375-1 · Cited by 2
A tri-modal contrastive learning framework for protein representation learning
Zhang, L., Guo, H., Schaffer, L. V., Ko, Y. S., Singh, D., Rahmani, H., Grotjahn, D. A., Villa, E., Gilson, M., Wang, W., et al. (2026) · Cell Reports Methods · doi:10.1016/j.crmeth.2026.101407 · Cited by 6
“22,23 Despite these successes, current PLMs predominantly focus on amino acid sequences while overlooking the wealth of complementary information available in other modalities.”
Chaudhary, S., Voigts, A., Bereket, M. D., Albert, M. L., Schwamborn, K., Zeggini, E., & Casale, F. P. (2026) · Genome biology · doi:10.1186/s13059-026-04031-z · Cited by 8
Descent from a common ancestor restricts exploration of protein sequence space
Isakova, L. H., Streltsova, E., Bochkareva, O., Vlasov, P. K., & Kondrashov, F. A. (2026) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2532018123 · Cited by 2
“When functions are experimentally tested they include mainly binding other molecules (proteins, nucleic acids, etc.) (5, 95), but even several examples of de novo enzyme design proved successful (6, 96).”
Routes towards an effective AI in CFD: an epistemological and technical perspective
Bauerheim, M. (2026) · Journal of Fluid Mechanics · doi:10.1017/jfm.2026.11256 · Cited by 4
Ding, K., Li, Z., Tu, T., Luo, J., & Luo, Y. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.02.03.703560 · Cited by 7
“DETANGO builds upon recent advances in pre-trained pLMs 36;37;39;43–46 , whose evolutionary plausibility predictions for protein mutants have been shown to correlate with proteins’ phenotypic readouts from MAVEs and the pathogenicity of human genetic variants.”
Chen, Y., Guo, X., & Xue, W. (2026) · Drug Discovery Today · doi:10.1016/j.drudis.2026.104634 · Cited by 7
Xu, M., Lu, M., Liu, P., Zhang, S., Chen, L., Liu, Q., Lin, Y., & Xie, L. (2025) · Briefings in Bioinformatics · doi:10.1093/bib/bbaf716 · Cited by 3
Multimodal diffusion for joint design of protein sequence and structure
Zhu, S., Gulati, S., Liu, Y., Kotnis, S., Sun, Q. H., & Shen, Y. (2025) · Protein Science · doi:10.1002/pro.70340 · Cited by 5
An interaction-derived graph learning framework for scoring protein–peptide complexes
Tao, H., Wang, X., & Huang, S. (2025) · Nature Machine Intelligence · doi:10.1038/s42256-025-01136-1 · Cited by 4
Lu, M.-Y., Hapeta, P., & Ledesma-Amaro, R. (2025) · Trends in Food Science & Technology · doi:10.1016/j.tifs.2025.105407 · Cited by 5
PepCCD: A Contrastive Conditioned Diffusion Framework for Target-Specific Peptide GenerationPreprint
Zhang, J., Zhou, Y., Zhu, T., & Zhu, Z. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.09.01.673427 · Cited by 1
“…in the effectiveness of protein sequence embeddings (Hayes et al. 2025; Brandes et al. 2022; Lin et al. 2023); and (2) strong adaptability to downstream tasks such as protein design after fine-tuning on task-specific data (Madani et al. 2023; Fer-ruz, Schmidt, and H ¨ ocker 2022; Lv et al. 2025).”
Talevi, A., Alberca, L. N., & Bellera, C. L. (2025) · Expert Opinion on Drug Discovery · doi:10.1080/17460441.2025.2555275 · Cited by 6
“For instance, Large Language Models (LLM) similar to ChatGPT (i.e. Chemical Language Models) are being actively used to assist molecular design [9–11].”
Scaling and Data Saturation in Protein Language Models
Spinner, A., Debenedictis, E., & Hudson, C. M. (2025) · doi:10.48550/arxiv.2507.22210 · Cited by 5
“…learning community with fantastic opportunities to grow the field and expand it into previously impossible questions, evidenced by all of the models trained on these collection of sequences (Jumper et al., 2021; Corn-man et al., 2024; Notin et al., 2022; Lin et al., 2023; Madani et al., 2023).”
Zhao, Z., Qu, Q., Sun, F., Zang, J., Zheng, B., Zhang, T., Zhao, G., Lv, C., & Wang, Z. (2025) · Biotechnology Advances · doi:10.1016/j.biotechadv.2025.108656 · Cited by 5
Protein design and optimization for synthetic cells
Frohn, B. P., Kohyama, S., & Schwille, P. (2025) · Nature Reviews Bioengineering · doi:10.1038/s44222-025-00318-1 · Cited by 7
“Autoregressive pLMs, which are trained to generate sequences without structural input, are a promising track in this regard 86,150,154,155 .”
Non-tissue culture genetic modifications for plant improvement
Li, X., Gao, Y.-H., Ma, H.-W., Wang, Y., Bu, T., Yin, W., Xia, X., & Wang, H. (2025) · Plant Molecular Biology · doi:10.1007/s11103-025-01594-6 · Cited by 8
“A protein language model, Pro-Gen, was developed, identifying an efficient Cas9 protein, OpenCRISPR-1, from more than 200 artificially designed CRISPR-Cas9 systems (Madani et al. 2023).”
Li, C.-F., Yan, Z., Ge, F., Yu, X., Zhang, J., Zhang, M., & Yu, D.-J. (2025) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.5c00478 · Cited by 5
CMADiff: Cross-Modal Aligned Diffusion for Controllable Protein Generation
Zhou, C., Qiu, Y.-X., Ling, T.-T., Li, J.-F., Liu, S.-H., Wang, X.-J., Song, J., & Xiang, W.-S. (2025) · arXiv.org · doi:10.48550/arxiv.2503.21450 · Cited by 5
“Exploring this sequence space through the design of proteins to produce biologically functional proteins remains a daunting challenge [3].”
Zhong, J., Zou, Z., Qiu, J., & Wang, S. (2025) · Briefings in Bioinformatics · doi:10.1093/bib/bbaf156 · Cited by 5
“[11] demonstrated the capabilities of large language models in generating functional protein sequences, while Watson et al .”
Khatri, K., Levy, R. M., & Haldane, A. (2025) · Physical Review Research · doi:10.1103/g5cx-1vhj · Cited by 3
AI for Science: Opportunities, Challenges, and Future Directions
Fu, V. (2026) · Journal of Data Science · doi:10.6339/25-jds1214 · Cited by 5
SeqProFT: Sequence-Only Protein Property Prediction With LoRA Finetuning
Zhang, S., & Liu, J. K. (2024) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2025.3636109 · Cited by 5
“With the latest developments in natural language processing, protein language models (PLMs) learn to understand the relationship between protein sequences, structures, and functions by treating amino acid sequences as text [2].”
Application of artificial intelligence in life science: Historical review and future perspectives
Shi, L., Wang, M., & Wang, X. (2024) · Fundamental Research · doi:10.1016/j.fmre.2024.10.014 · Cited by 6
“In recent years, the revolutionized advances of new AI methods, such as deep learning [ 25–27 ], large language model [ 28 ], transformer neural network [ 29 ] and generative AI [ 30 ], are expected to lead life science research into a new era.”
Towards Robust Evaluation of Protein Generative Models: A Systematic Analysis of MetricsPreprint
Страшнов, П. В., Shevtsov, A., Meshchaninov, V., Ivanova, M., Nikolaev, F., Kardymon, O. L., & Vetrov, D. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.10.25.620213 · Cited by 3
“…GANs, VAEs, and diffusion models, have been proposed and successfully applied to generate novel protein sequences Ferruz et al. (2022); Madani et al. (2023); Shin et al. (2021); Repecka et al. (2021); Sevgen et al. (2023); Lin & AlQuraishi (2023); Watson et al. (2023); Alamdari et al.…”
MethodWeinstein, E. N., Gollub, M. G., Slabodkin, A., Gardner, C. L., Dobbs, K., Cui, X.-B., Amin, A. N., Church, G. M., & Wood, E. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.09.13.612900 · Cited by 6
“;26;15;8;30 While generative models can produce astronomical numbers of novel sequences, and high-throughput experimental assays can evaluate billions of candidates or more, exact synthesis of individual sequences is limited by cost, and as a result most libraries do not exceed 10 5 candidates in…”
BackgroundAbGPT: De Novo Antibody Design via Generative Language ModelingPreprint
Kuan, D., & Farimani, A. B. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2409.06090 · Cited by 7
“16 20–29 The success of PLMs has led to the development of several foundation models that have significantly advanced our understanding of protein science.”
BackgroundMatsuzaka, Y., & Yashiro, R. (2024) · BioMedInformatics · doi:10.3390/biomedinformatics4030101 · Cited by 5
Generative β-hairpin design using a residue-based physicochemical property landscape
Satalkar, V. B., Degaga, G. D., Li, W., Pang, Y. T., McShan, A. C., Gumbart, J. C., Mitchell, J. C., & Torres, M. P. (2024) · Biophysical Journal · doi:10.1016/j.bpj.2024.01.029 · Cited by 4
“…big-data availability, researchers are shifting from the use of ML models, such as support vector machines (36–38) and random forests (39,40), for predictive and clustering studies toward more robust techniques such as attention-based deep learning (41,42) and natural language processing (43,44).”
Predictions from Deep Learning Propose Substantial Protein-Carbohydrate Interplay
Canner, S. W., Schnaar, R. L., & Gray, J. J. (2025) · doi:10.1101/2025.03.07.641884 · Cited by 4
“20 Further, we included the computationally designed and experimentally viable lysozymes from ProGen, 24 with structures predicted by the Colab distribution of AlphaFold2.”
Toward a Team of AI-made Scientists for Scientific Discovery from Gene Expression Data
Liu, H., Li, Y., Jian, J., Cheng, Y., Lu, J., Guo, S., Zhu, J., Zhang, M., Zhang, M., & Wang, H. (2024) · arXiv · doi:10.48550/arxiv.2402.12391 · Cited by 4
“Researchers have also been trying to incorporate LLMs into scientific discovery such as in chemistry (Bran et al., 2023; Guo et al., 2023), biotechnology (Madani et al., 2023) and medical (Singhal et al., 2023; Yang et al., 2023c) areas by training or finetuning LLMs on domain-specific data.”
BackgroundDeciphering enzymatic potential in metagenomic reads through DNA language models
Prabakaran, R., & Bromberg, Y. (2025) · Nucleic Acids Research · doi:10.1093/nar/gkaf836 · Cited by 4
“We note that much of the recent emphasis in the protein and nucleotide work with dLMs has focused on creating novel sequences that perform desired functions [ 32 , 33 ].”
Generating functional plasmid origins with OriGen
Irvine, J., Martinson, J. N. V., Arora, J., Hasham, S. I., Patel, J. R., Qi, E. T., Cress, B. F., & Rubin, B. E. (2025) · Nucleic Acids Research · doi:10.1093/nar/gkaf1198 · Cited by 4
“Protein generation models like ProGen [ 6 ] and ProtGPT2 [ 7 ] have demonstrated the ability to create novel functional proteins across a range of structural families.”
Roche, R., Tarafder, S., & Bhattacharya, D. (2024) · doi:10.1101/2024.07.27.605468 · Cited by 4
Emergence of Biological Structural Discovery in General-Purpose Language ModelsPreprint
Wang, L. (2026) · Research Square · doi:10.64898/2026.01.03.697478 · Cited by 3
“As we transition to instruction-tuned models, the intermediate-scale Llama 3.1-8B achieves 75% accuracy via zero-shot chat interaction, suggesting that general reasoning capabilities begin to bridge the modal gap.”
InfluentialFrom virtual experiments to biomedical insight with synthetic data
Victoriano, M., Pavlović, M., Sandve, G. K., Oliveira, H. P., Rocha, A., & Greiff, V. (2026) · Nature Machine Intelligence · doi:10.1038/s42256-026-01244-6 · Cited by 2
SwitchCraft: A Programmatic Framework for Designing State-Switching Proteins.
Jing, B., Bafna, M., Parsan, A., Ni, H. M., Kwabi-Addo, D., Bryson, B. D., Klivans, A., & Berger, B. (2026) · arXiv.org · doi:10.48550/arxiv.2605.31236 · Cited by 4
“When prompted with such attributes, PLMs have generated functional proteins such as lysozymes (Madani et al., 2023), tryptophan synthases (Lambert et al., 2025), green fluorescent proteins (Hayes et al., 2025) and even CRISPR nucleases (Ruffolo et al., 2025).”
Shen, H., Zhang, L., Ma, X., Yin, Y., Wang, J., & Tan, B. (2026) · Science China Life Sciences · doi:10.1007/s11427-025-3163-6 · Cited by 5
“At the protein level, autoregressive language models such as ProtGPT2 (Ferruz et al., 2022) and ProGen (Madani et al., 2023) can generate novel protein sequences with plausible function.”
Understanding Large Language Model Driven Social Bots: A Behavioral Analysis and Impact Assessment
Li, S., Yang, J., Zhao, K., & Jia, D. (2025) · IEEE Transactions on Computational Social Systems · doi:10.1109/tcss.2025.3605278 · Cited by 61
“Beyond conventional applications such as chatbots and search engines, researchers now explore integrating LLMs into traditional domains including education [20], healthcare [21], [22], scientific research [23], and programming [24] workflows.”
Artificial intelligence-assisted mining of polyethylene terephthalate hydrolases
Xiong, Q., Liu, G., Gao, S., Hao, Y., Li, C., Liu, J., Huang, J., Sinskey, A. J., Zhang, Y. X., Chen, X. H., et al. (2026) · New Biotechnology · doi:10.1016/j.nbt.2026.03.010 · Cited by 4
Zhang, J., Wang, S., Gao, L., & Li, Q. (2026) · Journal of Biotechnology · doi:10.1016/j.jbiotec.2026.03.012 · Cited by 4
Contrastive learning for antibody-antigen sequence-to-specificity predictionPreprint
Lee, H., Castro, K. M., Renwick, S., Stalder, L., Glänzer, W. S., Kumar, R. K., Chen, N., Scheck, A., Yermanos, A., Mason, D. M., et al. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.02.25.707916 · Cited by 3
Pawar, S., Mahajan, S., Sahu, P., & Bajaj, P. (2026) · Bioorganic chemistry (Print) · doi:10.1016/j.bioorg.2026.109620 · Cited by 4
Biotechnology meets AI: Accelerating next-generation of microbial breeding for food biomanufacturing
Tan, Y., Ni, H., Wu, T., Li, W., Zeng, H., & Wang, Y. (2025) · Food Research International · doi:10.1016/j.foodres.2025.118152 · Cited by 4
Consistent Synthetic Sequences Unlock Structural Diversity in Fully Atomistic De Novo Protein Design
Reidenbach, D., Cao, Z., Zhang, Z., Didi, K., Geffner, T., Zhou, G., Tang, J., Dallago, C., Vahdat, A., Kucukbenli, E., et al. (2025) · doi:10.48550/arxiv.2512.01976 · Cited by 4
“We assessed the in-silico co-designability of 𝒟 AFDB − clstr by folding its sequences (length ∈ [32,256]) with ESMFold and computing the C 𝛼 and all-atom RMSD between the folded and original AFDB structures.”
Protein-nucleic acid language model-assisted design of precise and compact adenine base editor
Ren, J., Yao, J.-W., Cao, Q., Li, Y.-N., Li, Y., Zhang, Z., Ge, X., Wang, S.-F., Zhang, Y., Wang, X.-G., et al. (2025) · Nature Communications · doi:10.1038/s41467-025-65311-z · Cited by 4
“Minor Revisions: In Fig.7, the authors delivered LNPs for in vivo editing of Pcsk9 in mice.”
Wang, K., Liu, M.-H., Zhang, Y., Li, W.-N., & Han, W.-W. (2025) · Journal of Agricultural and Food Chemistry · doi:10.1021/acs.jafc.5c05340 · Cited by 4
Deep learning‐based annotation of plant abiotic stress resistance genes for crops
Zhang, H., Liu, X., Liu, W., Wang, S., Li, Y., Xiang, W., Yang, Q., Zhang, A., Wang, G., Li, Y., et al. (2025) · The Plant Journal · doi:10.1111/tpj.70556 · Cited by 4
Sun, Z., Ting, Y.-S., Liang, Y., Duan, N., Huang, S., & Cai, Z. (2026) · The Astrophysical Journal Supplement Series · doi:10.3847/1538-4365/ae5d3a · Cited by 27
“…process faces distinct methodological challenges compared to fields like bio-chemistry (e.g., Jumper et al. 2021; Tunyasuvunakool et al. 2021; Madani et al. 2023; Notin et al. 2024) or materials research (Pyzer-Knapp et al. 2022; Merchant et al. 2023; Manica et al. 2023; Tran et al. 2024).”
Self-citationAnd… cut! – how conformational regulation of CRISPR-Cas effectors directs nuclease activity
Calvert, R. W., & Knott, G. J. (2025) · Biochemical Journal · doi:10.1042/bcj20240481 · Cited by 6
A Question Answering Dataset for Temporal-Sensitive Retrieval-Augmented Generation
Chen, Z., Min, E., Zhao, X. H., Li, Y., Jia, X., Liao, J., Li, J., Wang, S., Hu, B., & Yin, D. (2025) · Scientific Data · doi:10.1038/s41597-025-06098-y · Cited by 12
BioLangFusion: Multimodal Fusion of DNA, mRNA, and Protein Language Models
Mollaysa, A., Moskalev, A., Pati, P., Mansi, T., Prakash, M., & Liao, R. (2025) · arXiv.org · doi:10.48550/arxiv.2506.08936 · Cited by 4
“…sequences, models including ProtTrans (Elnaggar et al., 2021a), ESM family of models (Lin et al., 2022b), Ankh (Elnaggar et al., 2023), and ProGen2 (Madani et al., 2023) have achieved state-of-the-art results in structure prediction, function classification, and mutational fitness estimation via…”
DS-ProGen: A Dual-Structure Deep Language Model for Functional Protein Design
Li, Y.-T., Jiang, J.-Y., Wang, Z.-K., Lin, Z.-Q., He, D.-C., Shan, Y., Shao, Y., Li, J., Shi, X.-Y., Wang, J.-M., et al. (2025) · arXiv.org · doi:10.48550/arxiv.2505.12511 · Cited by 4
“Despite notable progress in applying language model approaches to protein science (Ferruz, Schmidt, and H¨ocker 2022; Madani et al. 2023; Nijkamp et al. 2023), most existing models for inverse folding rely heavily on singular structural representations either utilizing backbone geometric…”
Foundation Models for AI-Enabled Biological DesignPreprint
Moldwin, A., & Shehu, A. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.11610 · Cited by 5
“PLMs like Progen [21], Progen2 [22], and ProtGPT2 [23] also use standard architectures with minor enhancements.”
InfluentialComputer‐Aided Technology for Bioactive Protein Design and Clinical Application
Wang, C., Chen, Y., & Ren, L. (2025) · Macromolecular Bioscience · doi:10.1002/mabi.202500007 · Cited by 2
Zhang, C., Sun, Y., & Hu, P. (2025) · Journal of Cheminformatics · doi:10.1186/s13321-025-00979-5 · Cited by 5
“A protein language model operates by capturing intricate patterns and dependencies within protein sequences, enabling it to discern meaningful representations from amino acid sequences [14].”
DrugPred: An ensemble learning model based on ESM2 for predicting potential druggable proteins
Zhang, H.-Q., Liu, S.-H., Yu, J.-W., Li, R., Ye, D.-X., Jin, Y.-T., Huang, C.-B., & Deng, K.-J. (2025) · Future generations computer systems · doi:10.1016/j.future.2025.107801 · Cited by 4
AI-Empowered Genome Decoding: Applications of Large Language Models in Genomics
Li, S., Fan, W., & Zhou, Y. (2025) · Frontiers of digital education. · doi:10.1007/s44366-025-0051-1 · Cited by 4
Large Model Era: Deep Learning in Osteoporosis Drug Discovery
Xu, J., Wen, X., Sun, L. J., Xing, K., Xue, L., Zhou, S., Hu, J., Ai, Z., Kong, Q., Wen, Z., et al. (2025) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.4c02264 · Cited by 5
Deep Learning-Assisted Discovery of Protein Entangling Motifs
Deng, P., Xu, L., Wei, Y., Sun, F., Li, L., Zhang, W., & Gao, H. (2025) · Biomacromolecules · doi:10.1021/acs.biomac.4c01243 · Cited by 4
Can Hallucinations Help? Boosting LLMs for Drug Discovery
Yuan, S.-Z., Qu, Z., Kangen, A., & Farber, M. (2025) · doi:10.48550/arxiv.2501.13824 · Cited by 4
“Large language models (LLMs) have demonstrated impressive capabilities across diverse domains, including everyday applications (Ge et al. 2023; Yao et al. 2024), scientific discovery (Zhang et al. 2024b; Madani et al. 2023), and chemistry (Boiko et al. 2023; Zhang et al. 2024a).”
CogProg: Utilizing Large Language Models to Forecast In-the-Moment Health Assessment
Sprint, G., Schmitter‐Edgecombe, M., Weaver, R. H., Wiese, L. K., & Cook, D. J. (2024) · ACM Transactions on Computing for Healthcare · doi:10.1145/3709153 · Cited by 3
Generative AI Models for the Protein Scaffold Filling Problem
Qingge, L., Badal, K., Annan, R., Sturtz, J., Liu, X., & Zhu, B. (2024) · J. Comput. Biol. · doi:10.1089/cmb.2024.0510 · Cited by 4
“These models can effectively capture the intricate patterns and long-range dependencies present in biological sequences, enabling more accurate predictions and even de novo protein design (Madani et al., 2023).”
Roxas, R. E., & Recario, R. N. C. (2024) · Indonesian Journal of Electrical Engineering and Computer Science · doi:10.11591/ijeecs.v36.i1.pp252-263 · Cited by 4
“Domains include biotechnology in protein sequencing [41], spelling error detection and correction [38], clinical practice [42], education [43], public health [44], and code explanations [45].”
BackgroundLeveraging pretrained deep protein language model to predict peptide collision cross section
Nakai-Kasai, A., Ogata, K., Ishihama, Y., & Tanaka, T. (2024) · bioRxiv · doi:10.1038/s42004-025-01540-z · Cited by 4
“A number of deep protein language models for general purpose [25, 26, 27, 28, 29, 30, 31] and specific tasks [32, 33] have so far been proposed.”
Recent progress of activator of an industrially important enzyme–Nitrile hydratase
Sun, W.-F., Zhou, Z.-M., & Liu, Y. (2024) · Molecular Catalysis · doi:10.1016/j.mcat.2024.114307 · Cited by 4
An integrative approach to protein sequence design through multiobjective optimization
Hong, L., & Kortemme, T. (2024) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1011953 · Cited by 25
Li, H., Jiang, L., Yang, K.-X., Shang, S.-L., Li, M., & Lv, Z.-B. (2024) · International Journal of Molecular Sciences · doi:10.3390/ijms25137049 · Cited by 4
Range-limited Heaps' law for functional DNA words in the human genome.
Li, W.-T., Almirantis, Y., & Provata, A. (2024) · Journal of Theoretical Biology · doi:10.1016/j.jtbi.2024.111878 · Cited by 4
“…et al., 2007), such as bidirectional encoder representations from transformers (BERT) (Devlin et al., 2018) or generative pre-trained transformer (GPT) (Radford et al., 2019), have been applied to biomolecular sequences, in particular protein sequences (Rao et al., 2020; Madani et al., 2023).”
Decoding complexity: how machine learning is redefining scientific discovery
Vinuesa, R., Cinnella, P., Rabault, J., Azizpour, H., Bauer, S., Brunton, B. W., Elofsson, A., Jarlebring, E., Kjellström, H., Markidis, S., et al. (2024) · doi:10.48550/arxiv.2405.04161 · Cited by 4
NeuroFold: A Multimodal Approach to Generating Novel Protein Variants in silicoPreprint
Amani, K., Fish, M., Smith, M. D., & Castroverde, C. D. M. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.03.12.584504 · Cited by 1
“The final dataset consisted of 116,321 individual protein sequences from 38 protein families/taxa (Johnson et al., 2023; Madani et al., 2023; Notin et al., 2022; Rao, 2019; Repecka et al., 2021; Rocklin et al., 2017; Russ et al., 2020), including variants with single point mutations, multiple point…”
BackgroundMethodTpgen: a language model for stable protein design with a specific topology structure
Min, X., Yang, C., Xie, J. H., Huang, Y., Liu, N., Jin, X., Wang, T., Kong, Z., Lu, X., Ge, S., et al. (2024) · BMC Bioinformatics · doi:10.1186/s12859-024-05637-5 · Cited by 2
“ProGen [18], RITA [19], and ProtGPT2 [20] are expansive generative models trained with extensive datasets using the transformer decoder.”
BackgroundPredictions from deep learning propose substantial protein–carbohydrate interplay
Canner, S. W., Schnaar, R. L., & Gray, J. J. (2026) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2523342123 · Cited by 4
Ab-initio amino acid sequence design from protein text description with ProtDAT
Guo, X.-Y., Li, Y.-F., Liu, Y., Pan, X., & Shen, H.-B. (2025) · Nature Communications · doi:10.1038/s41467-025-65562-w · Cited by 3
RAG-ESM: Improving pretrained protein language models via sequence retrieval
Sgarbossa, D., & Bitbol, A.-F. (2025) · doi:10.1101/2025.04.02.646805 · Cited by 3
“Conditioning on sequence provides an alternative to conditioning on control tags, e.g. gene ontology terms (Nijkamp et al., 2023; Madani et al., 2023), and to multimodal pLMs using structural and functional features (Hayes et al., 2024; Wang et al., 2025).”
Property guidance for protein sequence generative models with ProteinGuide
Xiong, J., Gaur, I., Lukarska, M., Nisonoff, H. M., Oltrogge, L. M., Savage, D. F., & Listgarten, J. (2026) · Nature Biotechnology · doi:10.1038/s41587-026-03207-z · Cited by 5
Applications and limitations of AI tools in enzyme design
Teijeiro-Juiz, R., Egeler, N., Jamróg, G., Ringel, M., Brück, T., Geronimo, B. D., & Loll, B. (2026) · Protein Science · doi:10.1002/pro.70698 · Cited by 2
From waste to wealth: AI-driven enzyme innovation for straw transformation
Sun, X., Zhou, L., Li, Y., Shen, Q., Wang, Y., Jia, B., Yao, Y., & Pu, Z. (2026) · Bioresource Technology · doi:10.1016/j.biortech.2026.134875 · Cited by 3
Xu, T., Yang, Y., Zhu, R., Lin, W., Li, J., Zheng, Y. J., Zhang, P., Zhang, G., Zhao, G., & Jiao, N. (2026) · Nature Computational Science · doi:10.1038/s43588-026-00983-1 · Cited by 6
“Generative architectures may also inform future synthetic biology efforts 43,44 , although such applications remain prospective and dependent on experimental validation.”
ProtoCycle: Reflective Tool-Augmented Planning for Text-Guided Protein DesignPreprint
Ge, Y., Zhao, G., Li, S., Cheng, Z., Zhao, Z., Xia, H., Ke, G., Zhang, L., Gao, Z., & Wang, Y. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2604.16896 · Cited by 1
“On one hand, protein language models such as ProGen (Madani et al., 2023) and the ESM family (Rives et al., 2021; Lin et al., 2023) are pretrained on large sequence corpora and can be adapted to design tasks.”
Wu, J., Wang, C., Zeng, Y., Song, Y.-Y., Xu, S., Nie, J., Wang, T.-X., Ma, J.-C., Feng, J., & Wu, L.-H. (2026) · Cell Reports · doi:10.1016/j.celrep.2026.117275 · Cited by 3
A Long-Context Generative Foundation Model Deciphers RNA Design PrinciplesPreprint
Huang, Y., Lv, G., Cheng, A., Xie, W., Chen, M., Ma, X., Huang, Y., Tang, Y., Shi, Q., Wang, J., et al. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.03.17.712398 · Cited by 1
Applications of machine learning in plant biotechnology
Ogra, I. O., Silué, Y., Fawole, O. A., Aremu, A. O., & Opara, U. L. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-01166-7 · Cited by 3
Conditioning Protein Generation via Hopfield Pattern Multiplicity
Varner, J. D. (2026) · arXiv.org · doi:10.48550/arxiv.2603.20115 · Cited by 3
“Deep generative models, including variational autoencoders [1], autoregressive language models [2], and diffusion models [3], learn sequence distributions from large databases and can produce diverse candidates, but require substantial training data and GPU resources.”
Manifold-based learning for high-throughput single-peanut phenotyping
Peng, W.-K., Lin, X., Deng, P., Jiang, J., Ding, C., & Yuan, S.-J. (2026) · npj Systems Biology and Applications · doi:10.1038/s41540-026-00688-1 · Cited by 3
AI-enabled protein design facilitates future plant research and crop breeding
Lou, Y., Wu, T., Xia, F., Zhao, A., & Wang, X. (2026) · PLANT PHYSIOLOGY · doi:10.1093/plphys/kiag147 · Cited by 4
Manufacturing-aware generative models enable petascale synthesis of designed DNA
Weinstein, E. N., Gollub, M. G., Slabodkin, A., Dobbs, K., Cui, X.-B., Gardner, C. L., Grant, R. J., Gurung, K., Bailey, A., Amin, A. N., et al. (2026) · Nature Biotechnology · doi:10.1038/s41587-026-03020-8 · Cited by 3
Training-Free Generation of Protein Sequences from Small Family Alignments via Stochastic AttentionPreprint
Varner, J. D. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2603.14717 · Cited by 2
“We ran stochastic attention on all eight Pfam families (Table 1) and observed that the method consistently generated novel protein sequences while preserving family-level amino acid statistics, across a ten-fold range of family sizes and a seven-fold range of sequence lengths (Fig.”
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
