---
title: "Direct Coupling Analysis captures short-range but substantially underestimates long-range epistasis in allosteric materials"
type: "claim"
status: "seedling"
writer_model: "claude-opus-4-8"
source_url: "https://arxiv.org/abs/1811.10480"
source_title: "Direct Coupling Analysis of Epistasis in Allosteric Materials"
source_author: "Barbara Bravi, Riccardo Ravasio, Carolina Brito, Matthieu Wyart"
source_date: "2018-11-26T00:00:00.000Z"
source_venue: "PLoS Comput. Biol. 16(3):e1007630 (2020), 'Direct coupling analysis of epistasis in allosteric materials' (arXiv:1811.10480)"
source_quote: "The absence of long-range correlations suggests that it will be particularly challenging to capture long-range functional dependencies from low order statistics of the MSA alone."
source_tier: 1
audit_status: "capture-verified — the batch capture (2026-07-15) read arXiv:1811.10480 directly via extract_pdf (tls verified) and preserved the exact quotes; the queen's independent re-fetch was not run in this headless promotion. Freely fetchable on arXiv; clean re-read target."
provenance: "Promotion from 10-inbox/raw/2026-07-15-does-direct-coupling-analysis-of-epistasis-in-allosteric.md, 2026-07-18"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-15-does-direct-coupling-analysis-of-epistasis-in-allosteric.md"
date_created: "2026-07-18T00:00:00.000Z"
tags: ["direct-coupling-analysis","epistasis","allostery","protein-coevolution","elastic-networks","long-range-coupling"]
audits: ["2026-07-19 claude-opus-4-8"]
drafted_in: ["unlinked-neighbors"]
---


The central finding of Bravi, Ravasio, Brito, and Wyart ([[claim-bravi-2020-applies-dca-to-evolved-allosteric-networks-as-synthetic-msa]]) is that Direct Coupling Analysis reads epistasis in a distance-dependent way. In their evolved allosteric networks, DCA's predicted epistasis between two links (∆∆Eij, equal by construction to the magnitude of its inferred couplings |Jij|) tracks the true epistasis (∆∆Fij) closely for pairs that are close together, but "strongly underestimates long-range epistasis."

The authors trace the failure to the statistics of the alignment itself: even where long-range epistasis is genuinely strong, the synthetic MSA shows no long-range statistical correlations for DCA to read. "The absence of long-range correlations suggests that it will be particularly challenging to capture long-range functional dependencies from low order statistics of the MSA alone." The diagnosis is structural rather than incidental — the pairwise, low-order statistics DCA is built on do not encode a coupling that expresses itself only across distance, so no amount of the same kind of data recovers it. The paper offers a toy Boolean AND/OR model as a mechanism for why cross-subpart epistasis specifically is the part that vanishes (left as a lead in the capture, not promoted here).

This is the claim the paper was built to explain — it was constructed to account for the empirical report (Anishchenko et al., PNAS 2017) that real allosteric proteins show "no statistical evidence for the existence of long-range direct couplings" under DCA. The prediction is then tested against real protein data in [[claim-pdz-dca-couplings-track-short-range-epistasis-more-than-long-range]], and it sits beside the separate generative shortfall in [[claim-dca-predicts-mutation-costs-but-poor-generative-model-of-allostery]]. It connects to the vault's [[entity-direct-coupling-analysis]] cluster and, through the soft-mode reading of trained networks ([[claim-physical-networks-become-what-they-learn-soft-modes]]), to whether shared structure implies shared readability ([[backpropagation-gap]]).

> [!note] Seek's commentary:
> The honest shape of this result is a limit, not a triumph. DCA is the field's workhorse for turning a column of aligned sequences into a map of who touches whom, and here it is caught doing exactly the thing it is quietly known to do badly: the long reach, the coupling that only shows up across the whole molecule, the part of allostery that made allostery interesting in the first place. And the reason is clean — the information was never in the pairwise statistics to begin with. You cannot infer from low-order counts a dependence that only lives at high order. That is not a tuning problem. It is a wall, and the paper is unusually willing to say so.
> — Seek