EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Limitations of the refolding pipeline for de novo protein design

Kerlen T Korbeld iD, Vsevolod Viliuga iD, Maximilian J L J Fürst

Protein Science · 2026-06 · 35 (6)

DOI: 10.1002/pro.70613 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Protein Science, online erschienen am 06.05.2026. Verfasst von Kerlen T Korbeld und 2 weiteren Autor:innen. Beteiligte Einrichtung: Molecular Enzymology Group, University of Groningen, Groningen, The Netherlands. Hauptthemen laut MEDLINE: Proteins, Protein Refolding, Protein Engineering. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 72 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Netherlands Organization for Scientific Research NWO, European Union, SURF Cooperative.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

With the emergence of powerful deep learning‐based tools, computational protein design has become a widely accessible technique. Nowadays, it is possible to perform both sequence and structure design in a matter of minutes, making the technology attractive to the broader scientific community. In protein design campaigns, one of the most common in silico strategies to evaluate how well a sequence encodes a target structure is the so‐called self‐consistency or refolding pipeline. In this approach, a structure prediction model is used to refold the designed sequence to probe whether it is compatible with the intended structure, and is evaluated via two metrics linked to experimental success: the confidence score of the predicted structure (predicted local distance difference test) and the self‐consistency root‐mean‐square deviation, which measures how closely the refolded structure matches the target. In this work, we systematically evaluate how different models and structure prediction settings impact these metrics, and to what extent they can be used to reliably filter sequence design candidates. We show that evolutionary information can obscure folding models' abilities to assess sequence‐structure compatibility, reducing the predictive performance of refolding metrics for experimental success, particularly for designs that share homology with natural sequences. We further highlight limitations of refolding metrics, including their sensitivity to structural features, such as flexibility. Our findings raise awareness of potential pitfalls in refolding‐based evaluation and support more informed use of these metrics in protein design campaigns.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

ProteinsProtein RefoldingProtein EngineeringModels, MolecularProtein FoldingProtein ConformationComputational BiologyAmino Acid Sequence

Stichwörter der Autor:innen

ESMfoldalphaFolddo novo designprotein designinverse foldingrefolding pipelineself‐consistency pipeline

Substanzen

Proteins

Autor:innen und Institutionen

  • Molecular Enzymology Group, University of Groningen, Groningen, The Netherlands.
  • Department of Biochemistry and Biophysics and Science for Life Laboratory, Stockholm University, Stockholm, Sweden.
  • Biomolecular Mechanics, Max Planck Institute for Polymer Research, Mainz, Germany.

ORCID-Profile: Kerlen T Korbeld, Vsevolod Viliuga

Förderung

  • Netherlands Organization for Scientific Research NWO
  • European Union
  • SURF Cooperative
  • Swedish Research Council
  • European Commission
  • Vetenskapsrådet

Zitieren

APA

Korbeld, K. T., Viliuga, V., & Fürst, M. J. L. J. (2026). Limitations of the refolding pipeline for de novo protein design. Protein Science, 35(6). https://doi.org/10.1002/pro.70613

Vancouver

Korbeld KT, Viliuga V, Fürst MJLJ. Limitations of the refolding pipeline for de novo protein design. Protein Science. 2026;35(6). doi:10.1002/pro.70613

BibTeX
@article{Korbeld2026,
  author = {Korbeld, Kerlen T and Viliuga, Vsevolod and Fürst, Maximilian J L J},
  title = {Limitations of the refolding pipeline for de novo protein design},
  journal = {Protein Science},
  year = {2026},
  volume = {35},
  number = {6},
  doi = {10.1002/pro.70613}
}
Kostenlos starten →