arXiv News

First-hand information for everyone

Site

AboutPrivacy PolicyTerms of Use

© 2026 arXiv News

arXiv News

EnglishJapanese

Switch language

EnglishJapanese
Loading account…

First-hand information for everyone

Back to front page
Quantitative BiologyEnglishPublished

Protein language models lose contact‑map accuracy after swapping amino‑acid names, unlike classic Potts models

November 16, 2026arXiv: 2611.16881v1

Researchers found that protein language models (pLMs) do not keep accurate contact predictions when every amino‑acid letter is consistently renamed. The test, a global bijective relabeling that shuffles the alphabet but keeps sequence alignments unchanged, showed that all checked pLMs collapse their contact maps, whereas classical pseudolikelihood Potts models and inverse‑covariance estimators remain perfectly invariant.

The study introduced this diagnostic framework to probe whether contact readouts depend only on column statistics—what classic models use—or on explicit residue identities. Multiple architectures (ESM‑2, MSA Transformer, AMPLIFY) were tested at sizes from 8 million to 3 billion parameters. After relabeling, approximately 0.05 of the 38 contacts predicted by ESM‑2 650 M overlapped with the original set.

Results suggest that pLMs rely on “alphabet anchoring,” i.e., they use explicit, canonical residue identities rather than pure coevolutionary statistics. Even after increasing scale, performance still drops under relabeling.

However, the evaluation uses nine short, single‑domain proteins from four Pfam families and only three pLM families. It does not cover multi‑domain proteins or fine‑tuned models, and it tested only one grouping scheme based on BLOSUM62.

These findings show a gap between how pLMs and classical coevolutionary models compute contacts, calling for more diagnostics to understand what protein language models learn.

Source paper

Protein Language Models Fail To Capture Contact Predictions Under Invariant Relabeling

Quantitative BiologyArtificial Intelligence
View on arXiv
Protein language models lose contact‑map accuracy after swapping amino‑acid names, unlike classic Potts models | arXiv News