EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Learning constraints in protein evolution beyond known substitution frequencies

Raphaël Bricout iD, Hugues Roest Crollius iD, Auguste Genovesio iD

Molecular Biology and Evolution · 03.11.2026 · 43 (11)

DOI: 10.1093/molbev/msag270 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Molecular Biology and Evolution, online erschienen am 04.11.2026. Verfasst von Raphaël Bricout und 2 weiteren Autor:innen. Beteiligte Einrichtung: École normale supérieure, Institut de Biologie de l'ENS (IBENS), CNRS, INSERM, Paris 75005, France. Hauptthemen laut MEDLINE: Evolution, Molecular, Proteins, Amino Acid Substitution, Deep Learning. Die Arbeit nennt 38 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: French Government, ANR, MEMO LIFE.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

During the evolution of protein sequences, amino acid substitutions follow broad patterns that are well established, such as those induced by the physico-chemical properties of residues or the position of sites along the sequence. While these deviations from random changes have been exploited for practical applications, such as to simulate protein evolution or to score protein alignments, they are too general to predict the precise evolution of individual protein sequences. To investigate whether protein sequences contain contextual information about their subsequent evolution beyond that captured by conventional substitution statistics and evolutionary conservation, we trained deep learning models based on an Evolutionary Scale Modeling (ESM2) encoder to predict the rate, the position, and the nature of the amino-acid substitutions in the evolution of ancestral sequences and compared their performance to simpler and explainable heuristics. Our results show that predicting the rate of amino acid changes is achieved equally well by deep learning and by using evolutionary information in outgroup sequences. This suggests that the protein sequence does not contain information that significantly improves the predictions for this task. On the opposite, we show that deep learning models predict the position of these substitutions more accurately than statistics-based baselines. Predictions of the nature of substitutions largely outperform baselines, with over 37% accuracy overall. We observe that rapidly evolving sequences are harder to predict, and substitutions preferentially occur between chemically similar amino acids. Interestingly, predictive information is predominantly contained within a local sequence context of 16 residues. These findings indicate that the nature of substitutions in protein evolution is partly governed by predictive rules that remain to be discovered through a more accurate modeling that could describe evolutionary trajectories based on sequence context alone.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Evolution, MolecularProteinsAmino Acid SubstitutionDeep LearningModels, GeneticAmino Acid Sequence

Stichwörter der Autor:innen

amino-acid sequencesdeep learningevolution

Substanzen

Proteins

Autor:innen und Institutionen

  • École normale supérieure, Institut de Biologie de l'ENS (IBENS), CNRS, INSERM, Paris 75005, France.

ORCID-Profile: Raphaël Bricout, Hugues Roest Crollius, Auguste Genovesio

Förderung

  • French Government
  • ANR
  • MEMO LIFE
  • PSL
  • HPC
  • IDRIS
  • GENCI

Zitieren

APA

Bricout, R., Roest Crollius, H., & Genovesio, A. (2026). Learning constraints in protein evolution beyond known substitution frequencies. Molecular Biology and Evolution, 43(11). https://doi.org/10.1093/molbev/msag270

Vancouver

Bricout R, Roest Crollius H, Genovesio A. Learning constraints in protein evolution beyond known substitution frequencies. Molecular Biology and Evolution. 2026;43(11). doi:10.1093/molbev/msag270

BibTeX
@article{Bricout2026,
  author = {Bricout, Raphaël and Roest Crollius, Hugues and Genovesio, Auguste},
  title = {Learning constraints in protein evolution beyond known substitution frequencies},
  journal = {Molecular Biology and Evolution},
  year = {2026},
  volume = {43},
  number = {11},
  doi = {10.1093/molbev/msag270}
}
Kostenlos starten →