The DOI matches its Crossref record; 109 of the 112 references it cites were found in a registry or an index, and 1 are not checked yet; 8 of the 132 sources its claims cite were checked for support.
The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins
Open accessAbstract
Artificial intelligence-driven advances in protein structure prediction in recent years have raised the question: has the protein structure-prediction problem been solved? Here, with a focus on nonglobular proteins, we highlight the many strengths and potential weaknesses of DeepMind's AlphaFold2 in the context of its biological and therapeutic applications. We summarize the subtleties associated with evaluation of AlphaFold2 model quality and reliability using the predicted local distance difference test (pLDDT) and predicted aligned error (PAE) values. We highlight various classes of proteins that AlphaFold2 can be applied to and the caveats involved. Concrete examples of how AlphaFold2 models can be integrated with experimental data in the form of small-angle X-ray scattering (SAXS), solution NMR, cryo-electron microscopy (cryo-EM) and X-ray diffraction are discussed. Finally, we highlight the need to move beyond structure prediction of rigid, static structural snapshots toward conformational ensembles and alternate biologically relevant states. The overarching theme is that careful consideration is due when using AlphaFold2-generated models to generate testable hypotheses and structural models, rather than treating predicted models as de facto ground truth structures.
Verified 2026-10-07 · 1 reference not checked yet
Overview
AI snapshot
Written from the paper’s full text.
TL;DR
This review examines AlphaFold2’s strengths and failure modes for proteins beyond rigid globular folds, stressing that confidence scores do not guarantee agreement with experimental structures. It argues AF2 is best used as a hypothesis-generating complement to experimental data rather than as definitive structural truth.
Key findings
- AF2 has broad coverage—modeling 98.5% of the human proteome, with over 200 million precomputed predictions in the AlphaFold Protein Structure Database—but accuracy varies by target class.
- pLDDT and PAE are useful but imperfect: low pLDDT (<70) or high PAE (>5 Å) signal caution, yet high-confidence models can still deviate from experiments. On average, AF2 predicts about 50% of residues with high confidence, and one cited estimate suggests ~10% of high-confidence residues differ from native backbone by more than 2 Å.
- Non-globular and dynamic systems are especially challenging: peptides were excluded from training, IDPs/IDRs are often conformational ensembles, membrane proteins are underrepresented in the PDB (<3%), and AF2 struggles with alternative conformations, ligand/cofactor effects, and domain placement. Heterodimeric complex success rates were reported at 51% and 63% in two benchmarks, with antibody complexes particularly difficult.
- Integrating AF2 with SAXS, NMR, X-ray, and cryo-EM can improve models; AF-guided molecular replacement solved 187 of 215 structures in one benchmark, though cryo-EM maps worse than 6 Å are not recommended for fitting.
Methods
- Narrative review/perspective, not a primary research study; no new experimental data or systematic meta-analysis are presented.
- Synthesizes published benchmarks, case examples, and prior literature on AF2 workflow, pLDDT/PAE interpretation, protein classes, experimental integration, and conformational ensembles.
- Uses selected examples—such as OSBP1, insulin, ChRmine, SecA, and beta-endorphin amyloid—to illustrate strengths and failure modes.
Limitations
- Stated AF2 limitations include no training or validation for mutation effects, training bias toward soluble/PDB folds, peptide exclusion, membrane-protein underrepresentation, poor handling of IDPs/IDRs and alternative conformations, and cases where high pLDDT/PAE still mislead; inter-domain PAE accuracy is less benchmarked.
- Evident from the text, not stated by the authors: as a narrative review, the selection of examples and benchmarks is not systematic, and the conclusions depend on the cited literature rather than new quantitative assessment.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
74 Not checked yet · 15 Own interpretation · 8 No basis given · 4 Own result · 3 Partly supported · 2 Supported · accuracy not yet measured
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Several groups have reported cases of AF2 models with low PAE values (< 5 Å) that deviated from experimental data62,63”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Evaluation metrics and model reliability
What the sources say
- Partly supportedcites [62] McCafferty, C. L., Pennington, E. L., Papoulas, O., Taylor, D. W. & Marcotte, E. M. Does AlphaFold2 model proteins’ intracellular conformations? An experimental test using cross-linking mass spectrometry of endogenous ciliary proteins. Commun. Biol. 6, 421 (2023). sourceExperimental — accuracy not yet measured
The Results and discussion report AF2 models that violated experimental cross-link constraints despite high pLDDT scores, supporting the point that confident predictions can disagree with experimental data. But pLDDT is not PAE: the paper does not identify a violating model with PAE <5 Å, and it reports no violations in the PAE range 0–3.5. Its findings therefore do not establish the cited threshold. (LLM, conf=0.84)
- Partly supportedcites [63] Motmaen, A. et al. Peptide-binding specificity prediction using fine-tuned protein structure prediction networks. Proc. Natl Acad. Sci. USA 120, e2216697120 (2023). sourceExperimental — accuracy not yet measured
The Results (page unavailable) report that PAE only partly distinguished binding from non-binding peptides and that AlphaFold tended to dock non-binding peptides in the MHC binding groove. This supports the broader point that AlphaFold models can conflict with experimental binding evidence, but the paper does not identify models with PAE <5 Å or report those models deviating from experimental structures. (LLM, conf=0.76)
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.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Moreover, in the absence of NMR resonance assignments, AF2 models can be used as structural templates towards automated assignment90.”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Integration of AF2 models with experimental data
What the sources say
- Supportedcites [90] Klukowski, P., Riek, R. & Güntert, P. Time-optimized protein NMR assignment with an integrative deep learning approach using AlphaFold and chemical shift prediction. Sci. Adv. 9, eadi9323 (2023). source
The Results section reports automated NMR assignment using AlphaFold structures as input, including structure-based backbone amide assignment. It recommends predicted structures for automated chemical shift assignment with ARTINA; prior resonance assignments are not required as input. (LLM, conf=0.93)
plainconditional- 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.
- Supported
In the paper
“modeling of 98.5% of the human proteome5”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Main
What the sources say
- Not checked yetcites [5] Tunyasuvunakool, K. et al. Highly accurate protein structure prediction for the human proteome. Nature 596, 590–596 (2021). source
plaindirect- 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.
None of the sources cited for this claim have been checked yet.
In the paper
“are the best and most accurate answer to the structure prediction problem to date”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Main
What the sources say
- Not checked yetcites [1] Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021). source
- Not checked yetcites [17] Akdel, M. et al. A structural biology community assessment of AlphaFold2 applications. Nat. Struct. Mol. Biol. 29, 1056–1067 (2022). source
- Not checked yetcites [18] Baek, M. et al. Accurate prediction of protein structures and interactions using a three-track neural network. Science 373, 871–876 (2021). source
- Not checked yetcites [19] Lin, Z. et al. Evolutionary-scale prediction of atomic-level protein structure with a language model. Science 379, 1123–1130 (2023). 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
“whether the modeled structure resembles the apo or holo form of the protein is not immediately clear from analysis of pLDDT scores or PAE graphs32”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [32] Saldaño, T. et al. Impact of protein conformational diversity on AlphaFold predictions. Bioinformatics 38, 2742–2748 (2022). 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
“Benchmarking AF2 against membrane proteins represents a challenge since the membrane environment, which includes lipids and other proteins, is not directly considered by current versions of AlphaFold38”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [38] Hegedűs, T., Geisler, M., Lukács, G. L. & Farkas, B. Ins and outs of AlphaFold2 transmembrane protein structure predictions. Cell. Mol. Life Sci. 79, 73 (2022). source
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
“NMR structural ensembles offer a unique validation metric to assess the accuracy of predicted AlphaFold2 models since AlphaFold2 was trained on a subset of the PDB that excluded NMR data1,27–30”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [1] Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021). source
- Not checked yetcites [27] Alderson, T. R., Pritišanac, I., Kolarić, Đ., Moses, A. M. & Forman-Kay, J. D. Systematic identification of conditionally folded intrinsically disordered regions by AlphaFold2. Proc. Natl Acad. Sci. USA 120, e2304302120 (2023). source
- Not checked yetcites [28] Fowler, N. J. & Williamson, M. P. The accuracy of protein structures in solution determined by AlphaFold and NMR. Structure 30, 925–933 (2022). source
- Not checked yetcites [29] Zweckstetter, M. NMR hawk‐eyed view of AlphaFold2 structures. Protein Sci. 30, 2333–2337 (2021). source
- Not checked yetcites [30] Tejero, R., Huang, Y. J., Ramelot, T. A. & Montelione, G. T. AlphaFold models of small proteins rival the accuracy of solution NMR Structures. Front. Mol. Biosci. 9, 877000 (2022). source
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.
In the paper
“Azzaz et al have shown the difficulty in AF2 in modeling membrane proteins due to “epigenetic” factors (i.e., lipid environment, co-receptor induced structural changes, post-translational modifications) that control protein structure beyond the amino acid sequence40”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [40] Azzaz, F., Yahi, N., Chahinian, H. & Fantini, J. The epigenetic dimension of protein structure is an intrinsic weakness of the AlphaFold program. Biomolecules 12, 1527 (2022). source
boosteddirect- 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
“deviations of AF2 models from experimental structures also occur when co-factors, prosthetic groups, or ligands induce structural changes, either locally or allosterically”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
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.
In the paper
“Fowler et al revealed that NMR ensembles can be more accurate than static AF2 models for dynamic proteins28”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [28] Fowler, N. J. & Williamson, M. P. The accuracy of protein structures in solution determined by AlphaFold and NMR. Structure 30, 925–933 (2022). 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.
- Probably not a substantial claim. 15
- It may state two separate claims as one. 28
- The quote only partly states this claim. 1
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“allow users to run a modified AF2 protocol on open access servers in minutes”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Main
What the sources say
- Partly supportedcites [8] Mirdita, M. et al. ColabFold: making protein folding accessible to all. Nat. Methods 19, 679–682 (2022). sourceExperimental — accuracy not yet measured
The Main section describes ColabFold as a modified, faster AlphaFold2 workflow available through Google Colaboratory, which it says is free to logged-in users. The Abstract reports a throughput of nearly 1,000 structures per day on a one-GPU server, but does not establish that users can complete a prediction on the public platform within minutes. The paper does not address OpenFold. (LLM, conf=0.86)
- Partly supportedcites [9] Ahdritz, G. et al. OpenFold: retraining AlphaFold2 yields new insights into its learning mechanisms and capacity for generalization. Nat. Methods https://doi.org/10.1038/s41592-024-02272-z (2024). sourceExperimental — accuracy not yet measured
The Results section says OpenFold is an open-source implementation of AlphaFold2 that runs up to four times faster during inference. But the paper does not say that OpenFold and ColabFold let users run a protocol on open-access servers within minutes; it also says OpenFold reproduces the AlphaFold2 model architecture without modifications that alter its computations. (LLM, conf=0.85)
plaindirect- 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.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Hegedűs et al benchmarked several membrane proteins not included in the original AF2 training set and concluded that on average AF2 predicts transmembrane proteins as well as soluble proteins38”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Supportedcites [38] Hegedűs, T., Geisler, M., Lukács, G. L. & Farkas, B. Ins and outs of AlphaFold2 transmembrane protein structure predictions. Cell. Mol. Life Sci. 79, 73 (2022). source
The Results section identifies three transmembrane structures suitable for blind AF2 predictions because they or their homologues were absent from the training set. The Discussion concludes that AF2 likely predicts transmembrane proteins as well as soluble proteins, based on the similar pLDDT distributions; it presents that comparison cautiously rather than as a definitive accuracy measurement. (LLM, conf=0.89)
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. It may state two separate claims as one.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Preliminary reports also suggest that AF2 may also be able to predict oligomeric states of proteins and amyloids17,57”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Partly supportedcites [17] Akdel, M. et al. A structural biology community assessment of AlphaFold2 applications. Nat. Struct. Mol. Biol. 29, 1056–1067 (2022). sourceExperimental — accuracy not yet measured
The Results section on protein complex structures (page not provided) reports that AF2’s top-scoring model matched the known oligomeric state for 71 of 87 proteins, supporting the cautious claim about proteins. The passages from our copy do not establish the same ability for amyloids. (LLM, conf=0.86)
- Partly supportedcites [57] Jeppesen, M. & André, I. Accurate prediction of protein assembly structure by combining AlphaFold and symmetrical docking. Nat. Commun. 14, 8283 (2023). sourceExperimental — accuracy not yet measured
The Discussion reports that AlphaFold and AlphaFold-Multimer can model protein subunits and small oligomers, supporting the protein-oligomer portion of the claim. The paper does not address amyloids, and its prediction of larger assemblies requires symmetric docking rather than AF2 alone. (LLM, conf=0.86)
hedgedconditional- 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.
- Partly supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“the PH, CC, and ORD domain structure are assigned very high to high predicted confidence, while the FFAT domain is predicted with very low confidence”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > AlphaFold2’s AI-driven revolution
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. The quote only partly states this claim.
In the paper
“current versions of AF2 do not use energy functions that seek to identify native-like protein conformations, unlike its competitor Rosetta13”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Main
What the sources say
- Not checked yetcites [13] Alford, R. F. et al. The Rosetta all-atom energy function for macromolecular modeling and design. J. Chem. Theory Comput. 13, 3031–3048 (2017). source
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.
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
“requires significant computational resources to run the software locally (up to 3 terabytes of disk space and a modern NVIDIA GPUs with gigabytes of memory)”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Main
plaindirect- 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.
None of the sources cited for this claim have been checked yet.
In the paper
“houses over 200 million pre-run AF2 predictions”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Main
What the sources say
- Not checked yetcites [6] Varadi, M. et al. AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models. Nucleic Acids Res. 50, D439–D444 (2022). source
- Not checked yetcites [7] Varadi, M. et al. AlphaFold Protein Structure Database in 2024: providing structure coverage for over 214 million protein sequences. Nucleic Acids Res. 52, D368–D375 (2024). source
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.
None of the sources cited for this claim have been checked yet.
In the paper
“enables accurate prediction of three-dimensional protein structures from only the target amino acid sequence”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Main
What the sources say
- Not checked yetcites [1] Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021). source
- Not checked yetcites [2] Bertoline, L. M. F., Lima, A. N., Krieger, J. E. & Teixeira, S. K. Before and after AlphaFold2: an overview of protein structure prediction. Front. Bioinform. 3, 1120370 (2023).
- Not checked yetcites [3] Perrakis, A. & Sixma, T. K. AI revolutions in biology. EMBO Rep. 22, e54046 (2021). source
- Not checked yetcites [4] Bouatta, N., Sorger, P. & AlQuraishi, M. Protein structure prediction by AlphaFold2: are attention and symmetries all you need? Acta Crystallogr. D 77, 982–991 (2021). source
plainconditional- 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
“Tsaban et al showed that AF2 can be adapted to accurately model peptide/protein complexes irrespective of peptide length, although the results seemed biased towards helical structures and peptides that do not undergo large structural rearrangements upon binding24”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [24] Tsaban, T. et al. Harnessing protein folding neural networks for peptide–protein docking. Nat. Commun. 13, 176 (2022). source
hedgedcorrelational- 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
“the predicted Local Distance Difference Test (pLDDT) score, which ranges from 0 to 100 with higher values assigned higher confidence in the model”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > AlphaFold2’s AI-driven revolution
What the sources say
- Not checked yetcites [1] Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021). source
- Not checked yetcites [22] Mariani, V., Biasini, M., Barbato, A. & Schwede, T. lDDT: a local superposition-free score for comparing protein structures and models using distance difference tests. Bioinformatics 29, 2722–2728 (2013). source
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.
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
“For cases i and iv, low confidence in pLDDT scores and PAE graphs alerts users to interpret structures with caution, which is not immediately clear in cases ii and iii.”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
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.
None of the sources cited for this claim have been checked yet.
In the paper
“membrane proteins represent less than 3% of total structures in the PDB15, meaning that the training set used by AF2 was highly biased towards soluble proteins39”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [15] Laurents, D. V. AlphaFold 2 and NMR spectroscopy: partners to understand protein structure, dynamics and function. Front. Mol. Biosci. 9, 906437 (2022). source
- Not checked yetcites [39] Topitsch, A., Schwede, T. & Pereira, J. Outer membrane β-barrel structure prediction through the lens of AlphaFold2. Proteins 92, 3–14 (2024). source
boosteddirect- 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
“the best ranked AF2 models (selected on the basis of high pLDDT score) often did not exhibit the lowest Cα root-mean-square deviation (RMSD) relative to the experimental structure, suggesting the pLDDT metric used by AF2 to assess protein models is not optimal for classification of peptide conformations25”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [25] McDonald, E. F., Jones, T., Plate, L., Meiler, J. & Gulsevin, A. Benchmarking AlphaFold2 on peptide structure prediction. Structure 31, 111–119 (2023). source
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
“AF2 predicts the structure of individual domains well but deviates in placement of domains relative to each other (Fig. 2e)”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
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.
None of the sources cited for this claim have been checked yet.
In the paper
“Peptide structure prediction poses additional challenges given that the benchmark set used to train AF2 excluded peptides, the difficulty in generating robust MSAs for short sequences, and observations that many peptides exist in solution as conformational ensembles rather than a single static conformation1,24,25”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [1] Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021). source
- Not checked yetcites [24] Tsaban, T. et al. Harnessing protein folding neural networks for peptide–protein docking. Nat. Commun. 13, 176 (2022). source
- Not checked yetcites [25] McDonald, E. F., Jones, T., Plate, L., Meiler, J. & Gulsevin, A. Benchmarking AlphaFold2 on peptide structure prediction. Structure 31, 111–119 (2023). source
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.
In the paper
“Several groups have suggested that the default AF2 pipeline has difficulty in modeling alternative conformations45”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [45] Lane, T. J. Protein structure prediction has reached the single-structure frontier. Nat. Methods 20, 170–173 (2023). source
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.
Probably not a substantial claim.
In the paper
“AF2 also generates a Predicted Aligned Error (PAE) matrix, which evaluates the relative orientation and position of different parts (i.e., domains) of the model6”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > AlphaFold2’s AI-driven revolution
What the sources say
- Not checked yetcites [6] Varadi, M. et al. AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models. Nucleic Acids Res. 50, D439–D444 (2022). source
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.
In the paper
“AF2 performs poorly for targets embedded in membrane thickness outside the 15–35 Å range as well as targets with novel features not commonly present in the PDB”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [38] Hegedűs, T., Geisler, M., Lukács, G. L. & Farkas, B. Ins and outs of AlphaFold2 transmembrane protein structure predictions. Cell. Mol. Life Sci. 79, 73 (2022). source
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.
In the paper
“while the AF2 model of the channelrhodopsin ChRmine captures its overall fold, the modeled N-terminal region and extracellular loops deviate from its experimental high-resolution cryogenic electron microscopy (cryo-EM) structure (Fig. 2f), likely due to ChRmine’s unique covalent Schiff base feature41”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [41] Kishi, K. E. et al. Structural basis for channel conduction in the pump-like channelrhodopsin ChRmine. Cell 185, 672–689 (2022). source
hedgedconditionalSingle 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
“AF2 models with trans-membrane region lengths corresponding to non-physiological membrane thickness values can exhibit very high pLDDT scores (high model confidence), suggesting pLDDT scores alone are not sufficient to select native membrane protein conformations”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [38] Hegedűs, T., Geisler, M., Lukács, G. L. & Farkas, B. Ins and outs of AlphaFold2 transmembrane protein structure predictions. Cell. Mol. Life Sci. 79, 73 (2022). source
hedgedcorrelational- 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.
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 NMR structure of Ca2+ bound polycystin 2 deviates from the AF2 model, potentially due to conformational changes upon calcium binding (Fig. 2c)”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
hedgedconditional- 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.
In the paper
“it is important to remember that the PAE values are only confidence estimates. Furthermore, the accuracy of PAE graphs for inter-domain prediction has not been as extensively benchmarked as for intra-domain contacts36”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [36] David, A., Islam, S., Tankhilevich, E. & Sternberg, M. J. E. The AlphaFold database of protein structures: a biologist’s guide. J. Mol. Biol. 434, 167336 (2022). 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.
It may state two separate claims as one.
In the paper
“the AF2 model of insulin deviates significantly from its experimental NMR structure (Fig. 2b), potentially due to the inability of AF2 to orient cysteine pairs for disulfide bond formation31”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [31] Thornton, J. M., Laskowski, R. A. & Borkakoti, N. AlphaFold heralds a data-driven revolution in biology and medicine. Nat. Med. 27, 1666–1669 (2021). source
hedgedconditionalSingle 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 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
“high pLDDT or low PAE metrics, indicating high confidence in the prediction, do not promise agreement with native protein conformations, but instead estimate a likelihood for local and global coordinate positions and/or orientations”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > AlphaFold2’s AI-driven revolution
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 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 deviations can be broadly categorized into cases with i) inaccurate secondary, tertiary, and/or quaternary structure in regions where AF2 predicts low to moderate confidence, ii) inaccurate structure in regions where AF2 predicts high model confidence, iii) correct backbone structure but incorrect fine details (i.e., side chain rotamer placements), and iv) correct backbone”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
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.
In the paper
“Bryant et al reported a 63% success rate for heterodimeric complexes56”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [56] Bryant, P., Pozzati, G. & Elofsson, A. Improved prediction of protein–protein interactions using AlphaFold2. Nat. Commun. 13, 1265 (2022).
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.
None of the sources cited for this claim have been checked yet.
In the paper
“several groups have shown that modifications of AF2 have the potential to generate models that significantly deviate from each other, allowing for sampling of conformational landscapes”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
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.
In the paper
“other reports have compared pLDDT scores with crystallographic B-factors to suggest AF2 confidence metrics are unable to provide direct information on local flexibility66”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Evaluation metrics and model reliability
What the sources say
- Not checked yetcites [66] Carugo, O. pLDDT values in AlphaFold2 protein models are unrelated to globular protein local flexibility. Crystals 13, 1560 (2023). source
hedged- 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
“if the “best” AF2 model exhibits many residues with low pLDDT (< 70) or high PAE values (> 5 Å), the likelihood that the backbone structure matches the native conformation is very low and the model cannot be reliably interpreted”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Evaluation metrics and model reliability
hedgedcorrelational- 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
“additional strategies such as the combination of Vc, Qr, X2free, and Rsas metrics have been proposed as more robust evaluation metrics75.”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Integration of AF2 models with experimental data
What the sources say
- Not checked yetcites [75] Rambo, R. P. & Tainer, J. A. Accurate assessment of mass, models and resolution by small-angle scattering. Nature 496, 477–481 (2013). source
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.
In the paper
“AlphaMissense, does not explicitly determine the structural effects of a mutation on a protein but provides the probability of a missense variant being pathogenic51”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
What the sources say
- Not checked yetcites [51] Cheng, J. et al. Accurate proteome-wide missense variant effect prediction with AlphaMissense. Science 381, eadg7492 (2023). source
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.
In the paper
“Standard molecular replacement strategies require 3D coordinates of a template/homologous structure and work best when the template is < 2 Å Cα RMSD from the target structure80.”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Integration of AF2 models with experimental data
What the sources say
- Not checked yetcites [80] Abergel, C. Molecular replacement: tricks and treats. Acta Crystallogr. D 69, 2167–2173 (2013). source
plaincorrelational- 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 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
“which may be due to a training bias on stable structures or an inability to extract signal from small mutations through MSAs”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
hedgedconditional- 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
“Some groups have used MD simulations to suggest that pLDDT and PAE metrics provide information on dynamics and disorder42,65”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Evaluation metrics and model reliability
What the sources say
- Not checked yetcites [42] Ruff, K. M. & Pappu, R. V. AlphaFold and implications for intrinsically disordered proteins. J. Mol. Biol. 433, 167208 (2021). source
- Not checked yetcites [65] Guo, H.-B. et al. AlphaFold2 models indicate that protein sequence determines both structure and dynamics. Sci. Rep. 12, 10696 (2022).
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.
None of the sources cited for this claim have been checked yet.
In the paper
“Slight deviations in tertiary structure are noted, such as for the nuclear transport factor 2 derived de novo designed protein NT-9, but the overall structure is well described (Fig. 2j)”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > A critical evaluation of AlphaFold2’s applications
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
“The standard implementation of AlphaFold performs well only in detecting a single structural snapshot (the “ground state” structure), likely due to the lack of a large set of redundant protein conformers in the training set16,32.”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Beyond static snapshots: ensembles and conformational landscapes
What the sources say
- Not checked yetcites [16] Chakravarty, D. & Porter, L. L. AlphaFold2 fails to predict protein fold switching. Protein Sci. 31, e4353 (2022). source
- Not checked yetcites [32] Saldaño, T. et al. Impact of protein conformational diversity on AlphaFold predictions. Bioinformatics 38, 2742–2748 (2022). source
hedgedconditional- 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.
None of the sources cited for this claim have been checked yet.
In the paper
“This is especially helpful for large biological assemblies where methyl side-chain labeling affords an increase in signal and resolution91,92.”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Integration of AF2 models with experimental data
What the sources say
- Not checked yetcites [91] McShan, A. C. Utility of methyl side chain probes for solution NMR studies of large proteins. J. Magn. Reson. Open 14–15, 100087 (2023).
- Not checked yetcites [92] Ruschak, A. M. & Kay, L. E. Methyl groups as probes of supra-molecular structure, dynamics and function. J. Biomol. NMR 46, 75–87 (2009).
plaindirect- 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
“Current implementations of AF2 can provide highly accurate working models for most rigid, well-folded globular proteins, but may have issues predicting other classes of proteins.”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Conclusions and outlook
hedgedconditional- 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.
None of the sources cited for this claim have been checked yet.
In the paper
“Reports suggest AF2 models should not be fit into cryo-EM maps with resolution greater than 6Å98,99.”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Integration of AF2 models with experimental data
What the sources say
- Not checked yetcites [98] Dai, X., Wu, L., Yoo, S. & Liu, Q. Integrating AlphaFold and deep learning for atomistic interpretation of cryo-EM maps. Brief. Bioinform. 24, bbad405 (2023).
- Not checked yetcites [99] Alshammari, M., He, J. & Wriggers, W. Appraisal of AlphaFold2-predicted models in cryo-EM map interpretation. Microsc. Microanal. 29, 977–978 (2023). source
hedged- 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.
None of the sources cited for this claim have been checked yet.
In the paper
“The use of experimental data, especially SAXS, NMR, and cryo-EM, has also been described to guide modeling of ensembles and alternative conformations71,82,108.”
§ The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins > Beyond static snapshots: ensembles and conformational landscapes
What the sources say
- Not checked yetcites [71] Brookes, E., Rocco, M., Vachette, P. & Trewhella, J. AlphaFold-predicted protein structures and small-angle X-ray scattering: insights from an extended examination of selected data in the Small-Angle Scattering Biological Data Bank. J. Appl. Crystallogr. 56, 910–926 (2023). source
- Not checked yetcites [82] Abdollahi, H., Prestegard, J. H. & Valafar, H. Computational modeling multiple conformational states of proteins with residual dipolar coupling data. Curr. Opin. Struct. Biol. 82, 102655 (2023). source
- Not checked yetcites [108] Ramelot, T. A., Tejero, R. & Montelione, G. T. Representing structures of the multiple conformational states of proteins. Curr. Opin. Struct. Biol. 83, 102703 (2023). source
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.
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?
In the citing paper
“Investigators built an AlphaFold2 model of WSB1, refined it with molecular simulation, and conducted virtual screening across millions of compounds.”
From Artificial Intelligence in Clinical Oncology: From Productivity Enhancement to Creative Discovery (2025) · cites this paper as [104]
In the citing paper
“Approximately 20 candidates were selected for synthesis and testing, yielding four lead molecules.”
From Artificial Intelligence in Clinical Oncology: From Productivity Enhancement to Creative Discovery (2025) · cites this paper as [104]
In the citing paper
“A second example concerns WSB1, a protein linked to metastasis through hypoxia-response pathways in lung and pancreatic cancers but historically underexplored as a drug target [104].”
From Artificial Intelligence in Clinical Oncology: From Productivity Enhancement to Creative Discovery (2025) · cites this paper as [104]
Citations
Cites
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 · cited 7×
“…with its deep learning algorithm that enables accurate prediction of three-dimensional (3D) protein structures from only the target amino acid sequence, potentially solving the half-a-century-old protein structure-prediction problem: how to predict 3D structures from only sequence information 1–4 .”
BackgroundResultInfluentialColabFold - Making protein folding accessible to all
Mirdita, M., Schütze, K., Moriwaki, Y., Heo, L., Ovchinnikov, S., & Steinegger, M. (2021) · doi:10.1101/2021.08.15.456425 · Cited by 9,391 · cited 2×
“For targets without known homologs, such as computationally designed proteins, increasing the Perspective https://doi.org/10.1038/s41589-024-01638-w number of recycling iterations can improve the quality of the prediction 8 .”
BackgroundInfluentialVaradi, M., Anyango, S., Deshpande, M., Nair, S., Natassia, C., Yordanova, G., Yuan, D., Stroe, O., Wood, G., Laydon, A., et al. (2022) · Nucleic Acids Research · doi:10.1093/nar/gkab1061 · Cited by 7,850 · cited 1×
“Several efforts have alleviated these limitations, including the Alpha-Fold Protein Structure Database 6,7 , which houses over 200 million pre-run AF2 predictions, as well as ColabFold 8 and OpenFold 9 , which allow users to run a modified AF2 protocol on open-access servers in minutes.”
BackgroundEvolutionary-scale prediction of atomic level protein structure with a language model
Lin, Z., Akin, H., Rao, R., Hie, B., Zhu, Z., Lu, W., Smetanin, N., Verkuil, R., Kabeli, O., Shmueli, Y., et al. (2022) · doi:10.1101/2022.07.20.500902 · Cited by 5,990
Accurate prediction of protein structures and interactions using a 3-track network
Baek, M., DiMaio, F., Anishchenko, I., Dauparas, J., Ovchinnikov, S., Lee, G. R., Wang, J., Cong, Q., Kinch, L. N., Schaeffer, R. D., et al. (2021) · doi:10.1101/2021.06.14.448402 · Cited by 5,549
Protein complex prediction with AlphaFold-Multimer
Evans, R., O’Neill, M., Pritzel, A., Antropova, N., Senior, A., Green, T., Žídek, A., Bates, R., Blackwell, S., Yim, J., et al. (2021) · doi:10.1101/2021.10.04.463034 · Cited by 3,808 · cited 1×
“When more than one input sequence is provided, AlphaFold-Multimer or AF2Complex is used 20,21 .”
Highly accurate protein structure prediction for the human proteome
Tunyasuvunakool, K., Adler, J., Wu, Z., Green, T., Zielinski, M., Žídek, A., Bridgland, A., Cowie, A., Meyer, C., Laydon, A., et al. (2021) · Nature · doi:10.1038/s41586-021-03828-1 · Cited by 3,025 · cited 6×
“The authors of AF2 note that, because it is trained on both apo and holo structures from the PDB, models may still be consistent with the expected structure in the presence of ligands or cofactors despite their absence in the AF2 workflow 1,5 .”
BackgroundInfluentialAccurate proteome-wide missense variant effect prediction with AlphaMissense
Cheng, J., Novati, G., Pan, J., Bycroft, C., Žemgulytė, A., Applebaum, T., Pritzel, A., Wong, L. H., Zielinski, M., Sargeant, T., et al. (2023) · Science · doi:10.1126/science.adg7492 · Cited by 2,163 · cited 2×
“A recent adaption of AlphaFold, AlphaMissense, does not explicitly determine the structural effects of a mutation on a protein but provides the probability of a missense variant being pathogenic 51 .”
BackgroundMethodVaradi, M., Bertoni, D., Magana, P., Paramval, U., Pidruchna, I., Radhakrishnan, M., Tsenkov, M., Nair, S., Mirdita, M., Yeo, J., et al. (2024) · Nucleic Acids Research · doi:10.1093/nar/gkad1011 · Cited by 2,052 · cited 1×
“Several efforts have alleviated these limitations, including the Alpha-Fold Protein Structure Database 6,7 , which houses over 200 million pre-run AF2 predictions, as well as ColabFold 8 and OpenFold 9 , which allow users to run a modified AF2 protocol on open-access servers in minutes.”
BackgroundThe Rosetta All-Atom Energy Function for Macromolecular Modeling and Design
Alford, R. F., Leaver-Fay, A., Jeliazkov, J. R., O’Meara, M. J., DiMaio, F. P., Park, H., Shapovalov, M. V., Renfrew, P. D., Mulligan, V. K., Kappel, K., et al. (2017) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.7b00125 · Cited by 1,532
Research works printed but not linked: 6
- Bertoline, L. M. F., Lima, A. N., Krieger, J. E. & Teixeira, S. K. Before and after AlphaFold2: an overview of protein structure prediction. Front. Bioinform. 3, 1120370 (2023).
- Zhao, B., Ghadermarzi, S. & Kurgan, L. Comparative evaluation of AlphaFold2 and disorder predictors for prediction of intrinsic disorder, disorder content and fully disordered proteins. Comput. Struct. Biotechnol. J. 21, 3248–3258 (2023).
- McShan, A. C. Utility of methyl side chain probes for solution NMR studies of large proteins. J. Magn. Reson. Open 14–15, 100087 (2023).
- Ruschak, A. M. & Kay, L. E. Methyl groups as probes of supra-molecular structure, dynamics and function. J. Biomol. NMR 46, 75–87 (2009).
- Dai, X., Wu, L., Yoo, S. & Liu, Q. Integrating AlphaFold and deep learning for atomistic interpretation of cryo-EM maps. Brief. Bioinform. 24, bbad405 (2023).
- Johansson-Åkhe, I. & Wallner, B. Improving peptide–protein docking with AlphaFold-Multimer using forced sampling. Front. Bioinform. 2, 959160 (2022).
Cited by
Sorted by citations
Protein structure prediction via deep learning: an in-depth review
Meng, Y., Zhang, Z., Zhou, C., Tang, X., Hu, X., Tian, G., Yang, J., & Yao, Y. (2025) · Frontiers in Pharmacology · doi:10.3389/fphar.2025.1498662 · Cited by 52
Lee, M., Schafer, J. W., Prabakaran, J., Chakravarty, D., Clore, M. F., & Porter, L. L. (2025) · Nature Communications · doi:10.1038/s41467-025-60759-5 · Cited by 38
“I believe the revised manuscript is now well-positioned for publication.”
Integrating AlphaFold pLDDT Scores into CABS-flex for enhanced protein flexibility simulations
Wróblewski, K., & Kmiecik, S. (2024) · Computational and Structural Biotechnology Journal · doi:10.1016/j.csbj.2024.11.047 · Cited by 30
“By using pLDDT scores to identify regions of low confidence, these regions can be targeted in MD simulations to explore potential alternative conformations or to better understand the dynamic behavior of flexible regions [2,8].”
AlphaFold and what is next: bridging functional, systems and structural biology
Szczepski, K., & Jaremko, Ł. (2025) · Expert Review of Proteomics · doi:10.1080/14789450.2025.2456046 · Cited by 25
“…regions display incorrect secondary, tertiary, or quaternary structures; (2) examples of backbone structure side chain rotamer or individual domain structures are accurate but misaligned relative to each other; and (3) instances where even high-confidence regions are still inaccurate [37].”
Structural basis for the dynamic regulation of mTORC1 by amino acids
Valenstein, M. L., Wranik, M., Lalgudi, P. V., Linde-Garelli, K. Y., Choi, Y., Chivukula, R. R., Sabatini, D. M., & Rogala, K. B. (2025) · Nature · doi:10.1038/s41586-025-09428-7 · Cited by 24
Liu, Z., Guo, X., Xu, Y., & Wu, J. (2025) · Journal of Agricultural and Food Chemistry · doi:10.1021/acs.jafc.4c08630 · Cited by 23
Large language models for bioinformatics
Ruan, W., Lyu, Y., Zhang, J., Cai, J., Shu, P., Yang, G., Lu, Y., Gao, S., Wang, Y., Wang, P., et al. (2025) · Quantitative Biology · doi:10.1002/qub2.70014 · Cited by 37
“The typical models include AlphaFold [258], AlphaFold 2 [259], AlphaFold 3 [185], Graph ‐ BERT [260], MARPPI [261].”
Sutherland, C. A., Stevens, D. M., Seong, K., Wei, W., & Krasileva, K. V. (2025) · The Plant Cell · doi:10.1093/plcell/koaf109 · Cited by 20
Protein Structural Phylogenetics
Puente-Lelièvre, C., Malik, A. J., & Douglas, J. (2025) · Genome Biology and Evolution · doi:10.1093/gbe/evaf139 · Cited by 27
Ptaszek, A. L., Li, J., Konrat, R., Platzer, G., & Head-Gordon, T. L. (2024) · Journal of the American Chemical Society · doi:10.1021/jacs.4c10474 · Cited by 20
Fantini, J., Azzaz, F., Scala, C. D., Aulas, A., Chahinian, H., & Yahi, N. (2025) · Pharmacology & Therapeutics · doi:10.1016/j.pharmthera.2025.108797 · Cited by 20
Quantum computing and the implementation of precision medicine
Nassir, N., Hashmi, M., Raji, K. G., Jamalalail, B. A., Maksymowsky, A., Scherer, S. W., Alsheikh-Ali, A. A. A., & Uddin, M. J. (2025) · npj Genomic Medicine · doi:10.1038/s41525-025-00537-w · Cited by 24
“These challenges are exacerbated by memory limitations on classical computers, particularly when modeling large or complex proteins 46 .”
Dickey, S. W., Burgin, D. J., Antwi, A. N., Villaruz, A. E., Galac, M. R., Cheung, G. Y. C., Rostovtseva, T. K., Worrall, L. J., Lazarski, A. C., Cino, E. A., et al. (2025) · Nature Communications · doi:10.1038/s41467-025-62604-1 · Cited by 13
Subbanna, M. S., Winters, M. J., Örd, M., Davey, N. E., & Pryciak, P. M. (2025) · Journal of Biological Chemistry · doi:10.1016/j.jbc.2025.108225 · Cited by 16
AI/ML-empowered approaches for predicting T Cell-mediated immunity and beyond
Chao, C., Chiu, Y., Yeung, L. C., Yee, C., Jiang, C., & Shen, X. (2025) · Frontiers in Immunology · doi:10.3389/fimmu.2025.1651533 · Cited by 9
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
Physics-Grounded Evaluation to Guide Accurate Biomolecular PredictionPreprint
Lyu, N., Du, S., Shao, Q., Yang, Z. J., Ma, J., & Herschlag, D. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.06.30.662466 · Cited by 9
From Prompt to Drug: Toward Pharmaceutical Superintelligence
Zhavoronkov, A. A., Gennert, D. G., & Shi, J. (2026) · ACS Central Science · doi:10.1021/acscentsci.5c01473 · Cited by 15
Chakraborty, C., Bhattacharya, M., & Lee, S.-S. (2026) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2026.1739303 · Cited by 11
“In biological sequence databases, the generic “X” token typically represents an unknown or undetermined residue (Kemena and Notredame, 2009; Castro et al., 2023; Agarwal and McShan, 2024; Dotan et al., 2024; Testagrose and Boucher, 2025).”
Machine learning approaches for predicting protein-ligand binding sites from sequence data
Vural, O., & Jololian, L. (2025) · Frontiers in Bioinformatics · doi:10.3389/fbinf.2025.1520382 · Cited by 8
Protein Engineering Paving the Way for Next-Generation Therapies in Cancer
Naderiyan, Z., & Shoari, A. (2025) · International Journal of Translational Medicine · doi:10.3390/ijtm5030028 · Cited by 7
“These AI approaches complement experimental methods and may aid in the design of therapeutics when used in conjunction with empirical validation [79].”
Mahtha, S. K., Sureshkumar, V., & Mohanty, D. (2026) · NAR Genomics and Bioinformatics · doi:10.1093/nargab/lqag002 · Cited by 8
“Earlier benchmarking studies of AlphaFold2 against NMR structures have also observed significant differences between prediction results and solution conformations of proteins [ 23 , 44 , 45 ].”
De novo identification of protein domains in cryo-electron tomography maps from AlphaFold2 modelsPreprint
Lu, Y., Chen, G., Sun, F., Zhu, Y., & Zhang, Z. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.11.21.623534 · Cited by 9
“2a, the protein is partitioned into four well-defined domains consistent with the PAE plot: each domain shows high internal reliability, while interdomain orientations remain uncertain.”
InfluentialArtificial Intelligence in Clinical Oncology: From Productivity Enhancement to Creative Discovery
Kuno, M., Osumi, H., Udagawa, S., Yoshikawa, K., Ooki, A., Shinozaki, E., Ishikawa, T., Oba, J., Yamaguchi, K., & Sakurada, K. (2025) · Current Oncology · doi:10.3390/curroncol32110588 · Cited by 7
“A second example concerns WSB1, a protein linked to metastasis through hypoxia-response pathways in lung and pancreatic cancers but historically underexplored as a drug target [104].”
AlphaFold modeling uncovers global structural features of class I and class II fungal hydrophobins
Yang, L., Hicks, D. J., Russo, P. S., & McShan, A. C. (2025) · Protein Science · doi:10.1002/pro.70279 · Cited by 8
Self-citationTailored structured peptide design with a key-cutting machine approach
Leyva, Y. C., Torres, M. D. T., Oliva, C. A., de la Fuente‐Núñez, C., & Brizuela, C. A. (2025) · Nature Machine Intelligence · doi:10.1038/s42256-025-01119-2 · Cited by 9
TBM preferred to AlphaFold 3 for functional models of insect odorant receptors
Thanu, V. C., Jabeen, A., Zographos, S. E., Taylor, P. W., & Ranganathan, S. (2025) · Computational and Structural Biotechnology Journal · doi:10.1016/j.csbj.2025.08.028 · Cited by 7
“The RMSF values of the mutated residues in all 36 predicted complexes over the 100 ns MD simulations, averaged over three replicates. demonstrated remarkable accuracy often comparable to those obtained through experimental methods [69].”
Practical Outcomes From CASP16 for Users in Need of Biomolecular Structure Prediction
Abriata, L. A., & Peraro, M. D. (2025) · Proteins Structure Function and Bioinformatics · doi:10.1002/prot.70078 · Cited by 7
Beyond evolution: De novo designed protein toolkit rewriting the rules of synthetic biology
Zhang, G., Liu, C., Li, W., Lu, J., Li, A., & Zhu, L. (2025) · Biosafety and Health · doi:10.1016/j.bsheal.2025.09.004 · Cited by 6
Bias in the AlphaFold3 prediction of ligand-induced domain motion in enzymes
Yu, H., Bekar-Cesaretli, A. A., Lazou, M., Kozakov, D., Joseph‐McCarthy, D., & Vajda, S. (2026) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2530709123 · Cited by 7
“As machine learning methods move away from biophysics, it has become less clear what the program actually models (39).”
Matinja, A. I., Kamarudin, N. H. A., Leow, T. C., Oslan, S. N., & Ali, M. S. M. (2024) · Journal of Molecular Evolution · doi:10.1007/s00239-024-10219-3 · Cited by 6
Multiscale modeling of protofilament structures: A case study on insulin amyloid aggregates
Puławski, W., Koliński, A., & Koliński, M. (2024) · International Journal of Biological Macromolecules · doi:10.1016/j.ijbiomac.2024.138382 · Cited by 7
Vidya, F. N. U., Luo, Y., Wu, H., van der Donk, W. A., McShan, A. C., & Agarwal, V. (2025) · Nature Communications · doi:10.1038/s41467-025-64365-3 · Cited by 6
“While the models presented above are all in line with biochemical data, a doubt still lingers 47 .”
Self-citationEvaluating the 3D structure prediction tools to identify optimal MEBPVC structure models
Sahoo, P. S., & Burra, V. L. S. P. (2024) · Computational and Structural Biotechnology Reports · doi:10.1016/j.csbr.2024.100010 · Cited by 5
Advancing membrane-associated protein docking with improved sampling and scoring in Rosetta
Samnata, R., Harmalkar, A., Prathima, P., & Gray, J. J. (2024) · doi:10.1101/2024.07.09.602802 · Cited by 5
Thermal shift assay to identify ligands for bacterial sensor proteins
Monteagudo‐Cascales, E., Cano‐Muñoz, M., Genova, R., et al. (2025) · FEMS Microbiology Reviews · doi:10.1093/femsre/fuaf033 · Cited by 4
“However, despite the accuracy of AlphaFold models, they do not account for conformational flexibility or ligand-induced structural changes (Karelina et al. 2023 , Agarwal and McShan 2024 ).”
The arthropod P450 Enchiridion: An integrated web resource for research on P450s
Charamis, J., Dermauw, W., Leeuwen, T. V., et al. (2025) · Insect Biochemistry and Molecular Biology · doi:10.1016/j.ibmb.2025.104377 · Cited by 4
Bacterial Chemoreceptors Transmit Stimulus Signals Through Coupled Entropic Switches
Flack, C. E., & Parkinson, J. S. (2025) · Journal of Molecular Biology · doi:10.1016/j.jmb.2025.169614 · Cited by 4
Wani, A. K., Khan, R. M., Elboughdiri, N. A., Singh, J., Kriaa, K., Maatki, C., Hadrich, B., & Singh, R. (2026) · Food Bioscience · doi:10.1016/j.fbio.2026.108391 · Cited by 4
Feldman, J., Brogi, M., & Skolnick, J. (2026) · Computational and Structural Biotechnology Journal · doi:10.34133/csbj.0142 · Cited by 11
“AlphaFold’s performance degrades substantially on proteins with low sequence similarity to its training set [33, 34].”
Structure and mechanism of human vesicular polyamine transporter
Guo, Y. C., Yang, G., Liu, H., Chai, J. L., Chen, J., Shanklin, J., Liu, Q., Liu, B., & Lu, M. (2025) · Nature Communications · doi:10.1038/s41467-025-59549-w · Cited by 5
Sohn, M.-H., Kim, S., Jeong, H. J., Ko, I. Y., Moon, J. W., Lee, D., & Oh, J. (2024) · International Journal of Molecular Sciences · doi:10.3390/ijms26010137 · Cited by 4
Coskuner‐Weber, O., Gentili, P. L., & Uversky, V. N. (2025) · Biophysical Reviews · doi:10.1007/s12551-025-01286-x · Cited by 4
“AI-based structure prediction tools, such as AlphaFold2 and Alpha-Fold3, employ ML to predict the 3D structure of proteins or other biological molecules from their amino acid sequences (Agarwal and McShan 2024; Abramson et al. 2024).”
Zheng, S. (2025) · PLOS ONE · doi:10.1371/journal.pone.0313812 · Cited by 4
Feldman, J. D., Brogi, M., & Skolnick, J. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.02.25.708002 · Cited by 2
Evolution of macromolecular crystallography beamlines at the Swiss Light Source and SwissFEL
Wang, M. (2025) · Journal of Synchrotron Radiation · doi:10.1107/s1600577525005016 · Cited by 5
The emergence of novel versus known three-dimensional structures from random sequences
Yang, R., Yang, H., Davydenko, A., Mawaldi, Z., Kormos, R. C., Myerscough, D., Wu, Y., & DeGrado, W. F. (2026) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2535076123 · Cited by 4
“32 Models were clustered using a greedy algorithm (see methods) with a 1.0 Å RMSD cutoff, which provided sufficient resolution to separate folds with qualitatively different motifs and φ/ψ angle distributions.”
InfluentialLi, Y., Du, G., Chen, J., Lv, X., & Liu, L. (2026) · Biotechnology Advances · doi:10.1016/j.biotechadv.2026.108850 · Cited by 4
Lou, F., Zhou, W., Tunc‐Ozdemir, M., Yang, J., Velazhahan, V., Tate, C. G., & Jones, A. M. (2024) · Biochemistry · doi:10.1021/acs.biochem.4c00361 · Cited by 3
Bai, T., Xie, J., Liu, B., & Liu, Y. (2025) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2025.3548940 · Cited by 3
“Notably, although RNA 3D structure is difficult to obtain directly, similar to protein structure predicted by AlphaFold2 [29], [30], biological structure language models for RNA 3D structure prediction are now available, such as RhoFold+ [31], RNA-FM [32], RNA-MSM [33] and trRosettaRNA [34].”
High-throughput prediction of peptide structural conformations with AlphaFold2Preprint
Ille, A. M., Markosian, C., Burley, S. K., Whitford, P. C., Onuchic, J. N., Pasqualini, R., & Arap, W. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.12.03.626727 · Cited by 3
“…structural groupings, including α-helical membrane-associated peptides (AH MP), α-helical soluble peptides (AH SL), β-hairpin peptides (BHPIN), disulfide-rich peptides (DSRP), mixed secondary structure membrane-associated peptides (MIX MP), and mixed secondary structure soluble peptides (MIX SL).”
Liu, J., Xiao, J., Su, X., & Wang, Y. (2025) · Briefings in Bioinformatics · doi:10.1093/bib/bbaf331 · Cited by 2
“AlphaFold2 accurately predicts tertiary structures from amino acid sequences [22] and is particularly well suited for handling sequences with low homology.”
AlphaFold Prediction of Protein–Protein Interactions in the Flaviviridae Proteomes
Surya, W., Goh, J., Ponniah, C., & Torres, J. (2025) · International Journal of Molecular Sciences · doi:10.3390/ijms262010159 · Cited by 2
“The general cases of pitfalls during evaluation of AF2 model quality using PAE values used herein have been recently discussed [107].”
Vieni, C., Rothstein, T. E., Johnson, S., Schuetz, A. N., Norgan, A. P., & Patel, R. (2025) · iScience · doi:10.1016/j.isci.2025.114077 · Cited by 3
“While protein predictions from AlphaFold have revolutionized structural biology, these predictions can deviate from experimental models on a global and local scale 109,110 and predicting the impact of multiple, simultaneous amino acid mutations on protein structure is challenging.”
Chimeric Approach to Identify Molecular Determinants of Nicotinic Acetylcholine Receptors
Sapkota, P., Akaberi, S. M., Ghimire, B., & Sharma, K. (2026) · International Journal of Molecular Sciences · doi:10.3390/ijms27021091 · Cited by 3
“Current prediction algorithms struggle to predict the effect of point mutations and posttranslational modifications, especially in membrane proteins, and do not explicitly model lipid–protein interactions, which is critical for nAChR TMD function [133,134].”
Understanding protein folding with machine learning models? The case of AlphaFold2
Schuster, A. (2026) · Synthese · doi:10.1007/s11229-025-05426-4 · Cited by 3
“This is in line with the analysis by Agarwal and McShan (2024), who conclude that “[e]ven in the face of an impressive display of accuracy, AF2 is still best used to complement and extend interpretation of exper - imental data.””
InfluentialSebastiano, M. R., Apprato, G., Francisco, S., Ermondi, G., & Caron, G. (2026) · Medicinal Research Reviews · doi:10.1002/med.70038 · Cited by 5
“In addition, the original AF3 paper reports a tendency toward “hallucination,” as to say that defined structural features can be artificially and erroneously assigned to Intrinsically Disordered Regions [58, 109].”
From foundation models to autonomous agents in biology
Huang, S., Lang, M., Chen, Z., Yang, C., Huang, X., Mohtashaminia, Z., & Peng, Y. (2026) · Genomics Communications · doi:10.48130/gcomm-0026-0005 · Cited by 4
Han, H., Zhang, Q., Chen, Y., Wang, Z.-X., Su, X.-Y., Zhu, Y., Wang, X., & Feng, Z.-X. (2026) · International Journal of Biological Macromolecules · doi:10.1016/j.ijbiomac.2026.152018 · Cited by 2
The accuracy of electrostatic interactions captured by AI protein structure prediction models
Makhatadze, G. I. (2026) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2609610123 · Cited by 2
Zheng, Z.-Y., Shen, Z.-H., Xie, G., et al. (2024) · International Journal of Biological Macromolecules · doi:10.1016/j.ijbiomac.2024.138511 · Cited by 2
Şulea, T. A., Martin, E. C., Bugeac, C. A., Bectaș, F. S., Iacob, A.-L., Spiridon, L. N., & Petrescu, A. (2025) · International Journal of Molecular Sciences · doi:10.3390/ijms26020500 · Cited by 2
Vigneswaran, A., Buschmann, T. A., & Latham, M. P. (2025) · Journal of Magnetic Resonance · doi:10.1016/j.jmr.2025.107865 · Cited by 2
Resurgence of magnetic resonance techniques in the era of AlphaFold
Skinner, C. E., Haynes, B. A., & Isaacson, R. L. (2025) · Biophysical Reviews · doi:10.1007/s12551-025-01329-3 · Cited by 2
“AlphaFold also sometimes overpredicts rigidity, a function of the PDB training set having been dominated by crystal structures (Agarwal and McShan 2024).”
Galeazzi, F. M., da Silva, G. M., Arantes, P. R., Varghese, I., Shukla, A., & Rubenstein, B. M. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.05.09.653048 · Cited by 2
“2 AlphaFold’s ability to rapidly predict protein structures has had significant implications for understanding protein function, 3 analyzing protein interactions, 4 and furthering the development of new therapeutics.”
Predicting reverse-bound peptide conformations in MHC Class II with PANDORA
Rademaker, D. T., Parizi, F. M., Vreeswijk, M., et al. (2025) · Frontiers in Immunology · doi:10.3389/fimmu.2025.1525576 · Cited by 1
“However, their strength also determines their weaknesses, as their limitations are based on the availability of known (or similar) allele complexes (15 – 17).”
AlphaFold 3 accurately models natural variants of Helicobacter pylori catalase KatA
Gómez, A. A., & Baylink, A. (2025) · Microbiology Spectrum · doi:10.1128/spectrum.00670-25 · Cited by 1
“Studies suggest that Alpha Fold’s predictive accuracy is system-dependent and heavily influenced by the quality of experimental data used during its training (18, 19).”
Structural flexibility of a recombinant intrinsically disordered LEA protein from Ramonda serbica
Pantelić, A., Ilina, T. V., Milić, D., Gradišar, H., Radosavljević, J., & Vidović, M. (2025) · Scientific Reports · doi:10.1038/s41598-025-18648-w · Cited by 3
“These results must be interpreted with caution, due to AlphaFold’s biased prediction towards helical structures and its low confidence in modeling disordered regions 43,44 .”
da Silva, V. M., Thureau, A., Madern, D., et al. (2025) · Biophysical Journal · doi:10.1016/j.bpj.2025.11.009 · Cited by 1
Computational insights into the natural phage endolysin linker landscape
Cremelie, E., Boulay, A., Vázquez, R., et al. (2026) · BioData Mining · doi:10.1186/s13040-026-00526-4 · Cited by 1
“Hence, while AlphaFold2 has been demonstrated to deliver reliable structures for globular proteins, predictions of modular proteins have their limitations and should hence be interpreted with caution [62, 63].”
Kim, H., Lee, J., Lee, S., et al. (2026) · Journal of Neurology · doi:10.1007/s00415-026-13731-2 · Cited by 1
“It also has inherent limitations in capturing dynamic processes such as channel gating, post-translational modifications, or interactions with auxiliary subunits that are critical for channel function [1, 2].”
Toward universal binder discovery: Advances in display, computational design and in vivo platforms
Yang, J. K., Pineda, E., & Dickinson, B. C. (2026) · Current Opinion in Chemical Biology · doi:10.1016/j.cbpa.2026.102663 · Cited by 2
Bailey, J., Spina, S. C., Hu, A., Phan, N., Getman, R. B., & Kimmel, B. R. (2026) · ACS Engineering Au · doi:10.1021/acsengineeringau.5c00099 · Cited by 3
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
Howard, J. B., Narayanasamy, A., & Lindert, S. (2026) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.6c00265 · Cited by 1
Rajan, H., Devipriya, S. P., & Bhat, S. G. (2026) · Journal of Hazardous Materials Plastics · doi:10.1016/j.hazmp.2026.100066 · Cited by 1
Yang, L., Gupta, S., Mullininx, L. N., & McShan, A. C. (2026) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.6c01457 · Cited by 1
Self-citationSingh, S., Singh, R., & Sharma, S. (2026) · BioData Mining · doi:10.1186/s13040-026-00577-7 · Cited by 2
“However, AI methods also have their own limitations and should be used strategically in such scenarios [74].”
AlphaFold-based peptide structure prediction: Opportunities, limitations, and future directions
Zhang, B., Zhu, J., Chen, H.-F., & Liu, L. (2026) · Biotechnology Advances · doi:10.1016/j.biotechadv.2026.108977 · Cited by 4
Iebed, D., Gökler, T., van Ingen, H., et al. (2024) · ChemBioChem · doi:10.1002/cbic.202400589 · Cited by 1
Yu, W., Yu, L., Li, T., Wang, Z., Du, R., & Ping, W. (2025) · Fermentation · doi:10.3390/fermentation11040196 · Cited by 1
“Under Under the tight control of this cascade, the regulation between DAC and glucansucrase is the key point of protein interaction [58]. tight control of this cascade, the regulation between DAC and glucansucrase is the point of protein interaction [58].”
Calugaru, K., Yu, E. Y., Huang, S., González-Rodríguez, N., Coloma, J., & Lue, N. F. (2025) · Nucleic Acids Research · doi:10.1093/nar/gkaf245 · Cited by 1
“One likely explanation is that because the CST-PP-DNA complex can undergo more conformational changes than CST-primase-DNA, the computational challenges for machine learning-based algorithms are greater [ 46 ].”
Molecular recognition and structural plasticity in amyloid–nucleic acid complexes
Kukreja, R., & Latham, M. P. (2025) · Journal of Structural Biology · doi:10.1016/j.jsb.2025.108233 · Cited by 1
“And although AlphaFold has known issues modeling amyloidogenic proteins, future advances will certainly bring machine learning strategies to bear on uncovering these conformations (Agarwal and McShan, 2024).”
(2024) · doi:10.1101/2024.10.30.621084 · Cited by 1
“For example, quantitative data from large numbers of peptides, including abundant examples of non-binders, can provide training information to improve computational predictions of protein complexes and their binding affinities [86-88].”
Structures and zinc ion transport pathways of the human SLC39A family of metal transportersPreprint
Wang, X., Lorenz, C. D., Högstrand, C., & Maret, W. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.08.09.669477 · Cited by 1
The Lysis Cassette of Jumbophage phiKZPreprint
Manohar, P., Wan, J., Ganser, G., et al. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.11.11.687635 · Cited by 1
“Current artificial intelligence software for structure prediction is problematic for use on membrane proteins (Abramson et al., 2024b; Agarwal & McShan, 2024; Niazi, 2025).”
Narayanasamy, A., Drake, Z. C., Turzo, S. B. A., Rolland, A. D., Prell, J. S., Wysocki, V. H., & Lindert, S. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.02.16.706193 · Cited by 1
“…deep learning methods face limitations when predicting the structures of certain types of proteins, such as protein complexes or membrane-bound proteins, proteins with few known homologous sequences, intrinsically disordered proteins, proteins undergoing dynamic motions, 26 and ternary complexes.”
t2pmhc: A Structure-Informed Graph Neural Network to predict TCR-pMHC BindingPreprint
Polster, M., Stadelmaier, J., Ball, E., Scheid, J., Bauer, J., Nelde, A., Claassen, M., Dubbelaar, M. L., Walz, J. S., & Nahnsen, S. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.02.27.708137 · Cited by 1
Repetitive proteins that undergo large conformational changes evade structural prediction algorithms
Chang, M. P., Jin, T., Gudinas, A. P., Fernández, D., Alexander‐Katz, A., Matsui, T., & Mai, D. J. (2025) · The Journal of Chemical Physics · doi:10.1063/5.0304777 · Cited by 2
., A., Lata, S., & Chaudhuri, B. N. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.06.25.600654
Zheng, S. (2024) · doi:10.1101/2024.11.03.621778 · Cited by 4
“This innovative method has sparked considerable interest for its potential in accurately spotting disordered areas within proteins [26, 27].”
Calugaru, K., Yu, E. Y., González-Rodríguez, N., Coloma, J., & Lue, N. F. (2024) · doi:10.1101/2024.12.20.629258 · Cited by 1
Bressanelli, S., Fieulaine, S., & Tubiana, T. (2025) · Virology · doi:10.1016/j.virol.2025.110546
AlphaFold 3 accurately models natural variants of Helicobacter pylori catalase KatAPreprint
Perkins, A. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.06.02.657526 · Cited by 1
“Studies suggest that AlphaFold’s predictive accuracy is system-dependent and heavily influenced by the quality of experimental data used during its training 18,19 .”
Modeling CAPRI Targets of Round 55 by Combining AlphaFold and Docking
Singh, A., Copeland, M. M., Kundrotas, P. J., & Vakser, I. A. (2025) · Proteins Structure Function and Bioinformatics · doi:10.1002/prot.26853
The Impact of Neural Networks in Molecular and Cellular Biology
Guerroudji, W. H., Meddah, F. G., & Saadi, L. (2025) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-4571-0.ch003
Predicting blood group antigens: accessible or not?
Walser, P. J. (2025) · Blood · doi:10.1182/blood.2025029154
ELEN – Predicting Loop Quality in Protein Structure ModelsPreprint
Wieser, F., Kaltenbrunner, S., Zechner, A., Zach, M., Pock, T., & Oberdorfer, G. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.08.07.668890
Approaches to Study Proteins Encoded by Essential Genes
Cronan, J. E. (2025) · Proteins Structure Function and Bioinformatics · doi:10.1002/prot.70039
Metadata
- Authors
- Agarwal, V., & McShan, A. C.
- Affiliations
- Georgia Institute of Technology (US)
- Corresponding author
- Agarwal, Vinayak
- McShan, Andrew C.
- Venue
- Nature Chemical Biology
- Published
- 2024-06-21
- Volume
- 20
- Issue
- 8
- Pages
- 950-959
- ISSN
- 1552-4450 (ISSN-L)
- 1552-4469
- PMC
- PMC11956457
- OpenAlex
- W4399893066
- Cited by
- 155
- References
- 112
- Citations by year
- 2026: 79 · 2025: 63 · 2024: 13
- FWCI
- 6.31
- Type
- journal article
- Language
- en
- Open access
- green
- www.ncbi.nlm.nih.gov
- Licence
- Locations
- Nature Chemical Biology (publishedVersion)
- PubMed (publishedVersion)
- Europe PMC (PubMed Central) (acceptedVersion, other-oa)
- PubMed Central (acceptedVersion)
- Funders
- U.S. Department of Health & Human Services | NIH | National Institute of General Medical Sciences (R35GM142882)
- National Science Foundation (CHE-2238650)
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Structure and Dynamics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Structural Biology › Enzyme Structure and Function
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Machine Learning in Bioinformatics
- Keywords
- AlphaFold2, protein structure prediction, intrinsically disordered proteins, conformational ensembles, model reliability, pLDDT, small-angle X-ray scattering, solution NMR, cryo-electron microscopy, X-ray diffraction
- MeSH terms
- Artificial Intelligence, Humans, Models, Molecular, Protein Conformation, Proteins, X-Ray Diffraction, Protein Folding, Cryoelectron Microscopy, Scattering, Small Angle
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 2026-10-07, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-07 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
