The DOI matches its Crossref record; 61 of the 62 references it cites were found in a registry or an index; none of the 33 sources its claims cite is checked for support yet.
AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function
Open accessAbstract
AlphaFold 2 (AF2) was the star of CASP14, the last biannual structure prediction experiment. Using novel deep learning, AF2 predicted the structures of many difficult protein targets at or near experimental resolution. Here, we present our perspective of why AF2 works and show that it is a very sophisticated fold recognition algorithm that exploits the completeness of the library of single domain PDB structures. It has also learned local side chain packing rearrangements that enable it to refine proteins to high resolution. The benefits and limitations of its ability to predict the structures of many more proteins at or close to atomic detail are discussed.
Overview
AI snapshot
Written from the paper’s full text.
TL;DR
This Viewpoint argues that AlphaFold 2 (AF2) succeeds mainly as a sophisticated fold-recognition algorithm that exploits the near-completeness of the single-domain PDB structure library, plus learned local side-chain packing refinement. It also assesses AF2’s human-exome coverage and discusses implications and limits for inferring protein function, ligand binding, and phenotype.
Key findings
- AF2 often produced models with TM-score > 0.9, while MSA-based methods plateau around TM-score 0.85; TM-score > 0.85 indicates both global fold and details are correct.
- In human exome predictions, AF2 confidence score correlated with TM-score at 0.67. For low-quality models, only 10% had a closest template with TM-score ≥ 0.55; for good models, 53%; for high-accuracy models, 86%. These trends held with a current PDB library of almost 90,000 structures clustered at 70% sequence identity.
- Comparing 26,996 consensus human exome sequences, TASSER-VMT predicted at least one domain with TM-score > 0.4 for 89% of sequences; AF2 had confidence >70 for 67.4% and >60 for 86.9%. AF2 added 8.7% of human sequences with confidence >60, and the union provided confident models for at least one domain for 97.5% of human sequences.
- For ligand homology modeling, low-resolution structures are adequate, with at most a 20% improvement in enrichment factor going from lower resolution (TM-score > 0.4) to crystal structures, because binding sites are generally the most accurately predicted parts of a protein’s structure.
Methods
- This is a Viewpoint/perspective, not a new AF2 training study. It interprets AF2’s mechanism using prior deep-learning work and structural analyses.
- The authors used TM-align to generate structural alignments of AF2-predicted human proteins to a 41,000-member PDB library from 2016, and also examined results with a current PDB library of almost 90,000 structures clustered at 70% sequence identity.
- They analyzed correlations between AF2 confidence scores and TM-score/contact order, binned confidence as <70, 70–90, and >90, and compared AF2 coverage with TASSER-VMT on 26,996 consensus human exome sequences.
Limitations
- The paper states that structure alone does not provide biochemical function and cannot predict phenotypical function, which often depends on complex interactions among proteins, RNA, DNA, and metabolites.
- The paper states that AF2 predicts a particular conformational state; if an active conformation is predicted, it may not work for targeting the inactive state. It also notes that AF2 was trained on proteins with particular ions and might not accurately represent binding pockets under other conditions.
- The paper states that much work remains to establish AF2’s strengths and limitations, including whether it can predict effects of missense mutations, splice variants, chameleon sequences, cryptic pockets, and allosteric transitions. Evident from the text: the analysis is a perspective focused on the human exome and single-domain PDB completeness, and multimeric/multi-domain prediction is discussed as likely but not fully established.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
16 Not checked yet · 10 Own interpretation · 6 No basis given · 5 Own result
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
“For some multiple domain CASP14 targets, AF2 made high quality structure predictions.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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.
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
“Interestingly, the Pearson correlation coefficient of the average confidence score per protein with contact order is 0.28—a weak correlation of model quality with whether the fold has a sequential or nonsequential arrangement of secondary structural elements.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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.
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
“Having selected an approximate fold, AF2 sometimes refines the structure to experimental accuracy, with TM-scores greater than 0.9.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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: unclear — It is unclear why the paper says it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“These results are preserved when the current PDB library of almost 90,000 structures clustered at a sequence identity cutoff of 70% is used, with a shift toward higher TM-scores for the various prediction confidence levels.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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
“The correlation coefficient of AF2’s confidence score with TM-score is 0.67.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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
“then for low quality predicted structures only 10% of proteins have a closest template in the template library with a TM-score ≥ 0.55.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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.
In the paper
“There are 26,996 consensus human exome sequences between AF2 and a previously predicted set using TASSER-VMT.44”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
- Not checked yetcites [44] 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.
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
“Using the same domain partitioning as TASSER-VMT, AF2 provides an additional 8.7% of human sequences with a confidence score greater than 60.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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.
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 union of approaches provides confident models for at least one domain for 97.5% of human sequences.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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
“While AF2 generates significantly better structure predictions than existing methods on average, for the human exome, it is worthwhile to examine how many additional sequences have confident models compared to the prior art.44”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
- Not checked yetcites [44] 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.
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
“In contrast, AF2 learned information about the relationship of a sequence to its fold regardless of its contact order.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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.
None of the sources cited for this claim have been checked yet.
In the paper
“The key to why AF211 works is the fact the library of single domain protein structures is essentially complete.2,12,13”
- Also stated in: § Abstract · boosted
“However, by 2005, the library of single domain protein structures was shown to be essentially complete.2,12,13”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
boosted- 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.
- Also stated in: § Abstract · boosted
In the paper
“In 2005,2 we pointed out that if one had an algorithm that maps a protein’s sequence to the correct template structure in the existing PDB14 library (independent of their evolutionary relationship), the folding problem could be solved.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
- Not checked yetcites [2] source
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.
None of the sources cited for this claim have been checked yet.
In the paper
“Deep learning with residue covariation significantly improved structure prediction.18,19,23–25,33”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
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.
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
“Why was not fold recognition solved in 2005? The problem was the inability to match evolutionarily distant or unrelated proteins to their closest template PDB structure.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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.
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
“Its individual per residue networks learned local, residue specific packing details.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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: unclear — It is unclear why the paper says it holds.
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
“It has also learned local side chain packing rearrangements that enable it to refine proteins to high resolution.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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.
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
“Since AF2 recapitulates the trends of the SAdLSA self-alignment study, albeit with better model accuracy, it appears to have solved the fold recognition problem.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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.
In the paper
“Using novel deep learning, AF2 predicted the structures of many difficult protein targets at or near experimental resolution.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
- Not checked yetcites [1] 14th Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction. Protein Structure Prediction Center. https://predictioncenter.org/casp14/ (accessed September 2021).
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
“SAdLSA trained on purely α-helical proteins can align the sequences of distantly related pure β-proteins.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
- Not checked yetcites [35] 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
“As discussed in ref 37, deep learning’s ability to learn long-range correlations among residues enables it to map sequences to existing structures by exploiting PDB completeness for single domain proteins.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
- Not checked yetcites [37] 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.
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 an active conformation is predicted by AF2, then using it to target the inactive state will not work.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, the structure of a protein can be used to infer its biochemical function.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
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: unclear — It is unclear why the paper says it holds.
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 structures probably have a TM-score to the native ≥ 0.40.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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: unclear — It is unclear why the paper says it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“However, a protein can exist in one conformation in its “inactive” state and another in its “active” state.53–56”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
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
“TASSER-VMT predicts that at least one domain has a TM-score greater than 0.4 (indicating global fold similarity to the native structure) for 89% of sequences.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
- Not checked yetcites [44] 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. 11
- Probably not a substantial claim. 1
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 contrast, 53% of proteins with good models and 86% with high accuracy models have a TM-score ≥ 0.55.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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.
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 has a confidence score greater than 70 for 67.4% of sequences, while 86.9% have a confidence score greater than 60.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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.
None of the sources cited for this claim have been checked yet.
In the paper
“In practice, the structural space of protein–protein interfaces is small (<1000 structurally distinct interfaces),41,42 with nothing special about protein–protein interfaces.43”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
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 DESTI-NI convolutional neural network (CNN) deep learning approach to structure prediction23 demonstrated that deep learning learned the contact map patterns of secondary structure packing and eliminated nonphysical, isolated residue pairs.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
- Not checked yetcites [23] 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.
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
“Here, we present our perspective of why AF2 works and show that it is a very sophisticated fold recognition algorithm that exploits the completeness of the library of single domain PDB structures.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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: unclear — It is unclear why the paper says 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 reason that SAdLSA trained on helical proteins works for β-proteins is that the global topology of β-proteins is in the training set;2,12,13 the chain contour of helical proteins provides a globally correct fold for β-proteins.13”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
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 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
“First, structure alone does not provide the biochemical function of the protein nor can AF2 predict phenotypical function which is often the result of the complex interplay of proteins, RNA, DNA, and metabolites.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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.
None of the sources cited for this claim have been checked yet.
In the paper
“Contrary to popular belief that high resolution structures are required, using ligand homology modeling,45–48 low resolution structures are adequate, with at most a 20% improvement in enrichment factor on going from lower resolution (TM-score > 0.4) to crystal structures.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
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: Binding sites are generally the most accurately predicted parts of a protein's structure.
In the paper
“How inclusion of higher quality structures improves the ability to predict drug side effects, efficacy64 and the role of missense variations in non-Mendelian diseases depends on accurately characterizing the interplay of many molecular players.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
- Not checked yetcites [64] 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.
Probably not a substantial claim. 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
“Even though AF2 can accurately predict backbone and side chain conformations, it is for a particular conformational state.”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > 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: unclear — It is unclear why the paper says it holds.
It may state two separate claims as one.
In the paper
“Importantly, while knowledge of a protein’s structure alone does not provide its biochemical, much less, phenotypical effects, structure can be a component.64”
§ AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function > Abstract
What the sources say
- Not checked yetcites [64] 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.
The paper relies on this without support.
Binding sites are generally the most accurately predicted parts of a protein's structure.
“The reason is simple; binding sites are generally the most accurately predicted parts of a protein’s structure.”
The paper relies on this without support.
AF2 was trained on proteins having particular ions.
“Since AF2 was trained on proteins having particular ions, for example, it might not provide an accurate representation of the binding pocket under other conditions.”
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
“two key factors contributing to the success of AF2 are the completeness of the structural space of single-domain proteins and an abundance of sequences in sequence databases3.”
From AF2Complex predicts direct physical interactions in multimeric proteins with deep learning (2022) · cites this paper as [3]
In the citing paper
“Taken together, it is very likely that the neural network models trained for single-chain structure prediction have already learned the representations necessary to model protein complexes made of multiple single-chain proteins3.”
From AF2Complex predicts direct physical interactions in multimeric proteins with deep learning (2022) · cites this paper as [3]
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
I-TASSER: a unified platform for automated protein structure and function prediction
Roy, A., Küçükural, A., & Zhang, Y. (2010) · Nature Protocols · doi:10.1038/nprot.2010.5 · Cited by 6,212
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
I-TASSER server for protein 3D structure prediction
Zhang, Y. (2008) · BMC Bioinformatics · doi:10.1186/1471-2105-9-40 · Cited by 5,124
TM-align: a protein structure alignment algorithm based on the TM-score
Zhang, Y. (2005) · Nucleic Acids Research · doi:10.1093/nar/gki524 · Cited by 3,774
Improved protein structure prediction using potentials from deep learning
Senior, A. W., Evans, R., Jumper, J., Kirkpatrick, J., Sifre, L., Green, T., Qin, C., Žídek, A., Nelson, A. W. R., Bridgland, A., et al. (2020) · Nature · doi:10.1038/s41586-019-1923-7 · Cited by 3,378
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
A Method to Identify Protein Sequences That Fold into a Known Three-Dimensional Structure
Bowie, J. U., Lüthy, R., & Eisenberg, D. (1991) · Science · doi:10.1126/science.1853201 · Cited by 2,851
Database of homology‐derived protein structures and the structural meaning of sequence alignment
Sander, C., & Schneider, R. (1991) · Proteins Structure Function and Bioinformatics · doi:10.1002/prot.340090107 · Cited by 1,646
Improved protein structure prediction using predicted interresidue orientations
Yang, J., Anishchenko, I., Park, H., Peng, Z., Ovchinnikov, S., & Baker, D. (2020) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1914677117 · Cited by 1,421
Cited by
Sorted by citations
Before and after AlphaFold2: An overview of protein structure prediction
Bertoline, L. M. F., Lima, A. N., Krieger, J. E., & Teixeira, S. K. (2023) · Frontiers in Bioinformatics · doi:10.3389/fbinf.2023.1120370 · Cited by 330
AF2Complex predicts direct physical interactions in multimeric proteins with deep learning
Gao, M., Nakajima An, D., Parks, J. M., & Skolnick, J. (2022) · Nature Communications · doi:10.1038/s41467-022-29394-2 · Cited by 252
Self-citationProteins Plus : a comprehensive collection of web-based molecular modeling tools
Stierand, K., Diedrich, K., Ehrt, C., Flachsenberg, F., Graef, J., Sieg, J., Penner, P., Poppinga, M., Ungethüm, A., & Rarey, M. (2022) · Nucleic Acids Research · doi:10.1093/nar/gkac305 · Cited by 225
The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins
Agarwal, V., & McShan, A. C. (2024) · Nature Chemical Biology · doi:10.1038/s41589-024-01638-w · Cited by 155
Comprehensive Survey of Consensus Docking for High-Throughput Virtual Screening
Blanes‐Mira, C., Fernández-Aguado, P., de Andrés-López, J., Fernández-Carvajal, A. M., Ferrer-Montiel, A. V., & Fernandez-Ballester, G. J. (2022) · Molecules · doi:10.3390/molecules28010175 · Cited by 113
Metadata
- Authors
- Skolnick, J., Gao, M., Zhou, H., & Singh, S.
- Affiliations
- Georgia Institute of Technology (US)
- Twilight Design, 4 Adams Road, Kendall Park, New Jersey 08824, United States
- Corresponding author
- Skolnick, Jeffrey
- Venue
- Journal of Chemical Information and Modeling
- Publisher
- American Chemical Society (ACS)
- Published
- 2021-09-29
- Volume
- 61
- Issue
- 10
- Pages
- 4827-4831
- ISSN
- 1549-9596 (ISSN-L)
- 1549-960X
- PMC
- PMC8592092
- OpenAlex
- W3201869122
- Semantic Scholar
- 34ab2f845e0b6c1693a4fc92bf9593b32be0d803
- Cited by
- 251
- References
- 62
- Citations by year
- 2026: 28 · 2025: 58 · 2024: 64 · 2023: 66 · 2022: 31 · 2021: 2
- FWCI
- 12.86
- Type
- journal article
- Language
- en
- Open access
- green
- www.ncbi.nlm.nih.gov
- Licence
- Locations
- Journal of Chemical Information and Modeling (publishedVersion)
- PubMed (publishedVersion)
- PubMed Central (acceptedVersion)
- Funders
- U.S. Department of Health and Human Services (GM-118039)
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Structure and Dynamics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Machine Learning in Bioinformatics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Structural Biology › Enzyme Structure and Function
- Keywords
- AlphaFold2, protein structure prediction, CASP14, deep learning, protein fold recognition, side-chain packing, sequence-structure-function relationships
- MeSH terms
- Algorithms, Amino Acid Sequence, Proteins, Protein Folding
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
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
