Inter-residue, inter-protein and inter-family coevolution: bridging the scales
Pith reviewed 2026-05-25 16:57 UTC · model grok-4.3
The pith
Coevolution of interacting proteins spans residue, protein, and family scales that a single statistical framework can bridge using large sequence datasets.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
Interacting proteins coevolve at multiple but interconnected scales, from the residue-residue over the protein-protein up to the family-family level. The recent accumulation of enormous amounts of sequence data allows for the development of novel, data-driven computational approaches. Notably, these approaches can bridge scales within a single statistical framework.
What carries the argument
A single statistical framework that models coevolution simultaneously at residue-residue, protein-protein, and family-family scales.
If this is right
- Methods now used on isolated scales can be extended to treat residue, protein, and family levels inside one calculation.
- Evolutionary information extracted this way can supply constraints for structural modeling and interaction prediction.
- The framework supports an evolutionary informed structural systems biology that links molecular detail to cellular organization.
- Current scale-specific applications can serve as building blocks for multi-scale models.
Where Pith is reading between the lines
- Bridging the scales could reveal how a mutation at one residue position influences interaction specificity across an entire protein family.
- The same framework might be tested on well-characterized systems such as two-component signaling or metabolic pathways to check consistency across scales.
- Combining the coevolution statistics with experimental structure data could tighten the constraints available for modeling.
Load-bearing premise
The recent accumulation of enormous amounts of sequence data allows for the development of novel, data-driven computational approaches that can bridge scales.
What would settle it
A demonstration that no existing or foreseeable statistical model can recover accurate coevolutionary signals when applied simultaneously to residue pairs, protein pairs, and family pairs from the same interacting system would falsify the bridging claim.
Figures
read the original abstract
Interacting proteins coevolve at multiple but interconnected scales, from the residue-residue over the protein-protein up to the family-family level. The recent accumulation of enormous amounts of sequence data allows for the development of novel, data-driven computational approaches. Notably, these approaches can bridge scales within a single statistical framework. While being currently applied mostly to isolated problems on single scales, their immense potential for an evolutionary informed, structural systems biology is steadily emerging.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript is a perspective piece arguing that coevolutionary signals at residue, protein, and family scales can be bridged within a single statistical framework, enabled by the recent growth in sequence data, with emerging potential for evolutionary-informed structural systems biology. No new model, derivation, dataset, or cross-scale validation is presented.
Significance. The aspirational unification of multi-scale coevolution analysis could, if achieved, contribute to systems-level understanding of protein interactions. However, the manuscript provides only a high-level discussion of existing methods' potential without demonstrating a concrete bridging framework, quantitative bounds, or falsifiable predictions, so its immediate scientific impact remains limited.
major comments (1)
- Abstract and overall text: The central claim that 'these approaches can bridge scales within a single statistical framework' is stated as a notable capability but is not supported by any specific unified model, example calculation, or reference to a cross-scale implementation; the text remains at the level of suggesting future applications rather than showing how bridging is achieved.
Simulated Author's Rebuttal
We thank the referee for their review of our perspective manuscript. We address the major comment below, noting that this is a perspective article whose purpose is to outline conceptual opportunities rather than to deliver new empirical demonstrations.
read point-by-point responses
-
Referee: Abstract and overall text: The central claim that 'these approaches can bridge scales within a single statistical framework' is stated as a notable capability but is not supported by any specific unified model, example calculation, or reference to a cross-scale implementation; the text remains at the level of suggesting future applications rather than showing how bridging is achieved.
Authors: We agree that the manuscript does not present a new unified model, derivation, or cross-scale validation; as a perspective piece this is outside its scope. The claim refers to the fact that the same class of statistical models (e.g., inverse Potts or covariance-based methods) can be formulated identically whether the variables are residues, whole proteins, or protein families, and that the recent growth in sequence data now makes joint or hierarchical inference across these levels feasible in principle. The text reviews existing single-scale applications and cites early multi-scale efforts, but does not claim to have executed a concrete bridging implementation. We will revise the abstract and relevant sections to make this distinction clearer and to avoid any implication that a completed framework is already demonstrated. revision: partial
Circularity Check
No derivation chain; perspective piece only
full rationale
The paper is a perspective/discussion article. Its abstract and text contain no equations, models, derivations, fitted parameters, or quantitative predictions. The central claim is aspirational (sequence data enables bridging scales in one framework) and is not supported by any technical construction that could reduce to its own inputs. No self-citations, ansatzes, or uniqueness theorems are invoked in a load-bearing way. This is the expected outcome for a non-technical review.
Axiom & Free-Parameter Ledger
Reference graph
Works this paper leans on
-
[1]
Marks DS, Colwell LJ, Sheridan R, Hopf TA, Pagnani A, Zecchina R, Sander C: Protein 3D structure computed from evolutionary sequence variation. PLoS One 2011, 6:e28766. 18. Dago AE, Schug A, Procaccini A, Hoch JA, Weigt M, Szurmant H: Structural basis of histidine kinase autophosphorylation deduced by integrating genomics, molecular dynamics, and mutagene...
work page 2011
-
[2]
PLoS Comput Biol 2015, 11:e1004262
Malinverni D, Marsili S, Barducci A, De Los Rios P: Large-Scale Conformational Transitions and Dimerization Are Encoded in the Amino-Acid Sequences of Hsp70 Chaperones. PLoS Comput Biol 2015, 11:e1004262. ● The paper predicts the dimerization of Hsp70 chaperones on the bases of coevolutionary signals, which are not consistent with the monomeric protein st...
work page 2015
-
[3]
Methods Mol Biol 2012, 804:167-177
Pellegrini M: Using phylogenetic profiles to predict functional relationships. Methods Mol Biol 2012, 804:167-177. 45. Croce G, Gueudré T, Ruiz-Cuevas MV, Figliuzzi M, Szurmant H, Weigt M: A multi-scale coevolutionary approach to predict protein-protein interactions. in preparation 2017. 46. Jones DT, Singh T, Kosciolek T, Tetchner S: MetaPSICOV: combinin...
work page 2012
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.