Can AlphaFold2 predict protein-peptide complex structures accurately?
Abstract
AO_SCPLOWBSTRACTC_SCPLOWIn this preprint, we investigated whether AlphaFold2, AF2, can predict protein-peptide complex structures only with sequence information. We modeled the structures of 203 protein-peptide complexes from the PepBDB DB and 183 from the PepSet. The structures were modeling with concatenated sequences of receptors and peptides via poly-glycine linker. We found that for more than half of the test cases, AF2 predicted the bound structures of peptides with good accuracy, C-RMSD of a peptide < 3.0 [A]. For about 40% of cases, the peptide structures were modeled with an accuracy of C-RMSD < 2.0 [A]. Our benchmark results clearly show that AF2 has a great potential to be applied to various higher-order structure prediction tasks.
02 /Specification & Classification
Click any tag to browse matching papers- Citations
- 83
- Assay
- Research Focus
- Approach
Abstract summary
Claude Opus 5.5 · xhigh effort · 2 Nov 2026 · from the abstract · CC BY-NC-ND 4.0
Concatenating receptor and peptide sequences via a poly-glycine linker enables AlphaFold2 to predict bound peptide conformations within 3.0 A C-RMSD for over half of benchmarked protein-peptide complexes.
Whether AlphaFold2 (AF2) can predict protein-peptide complex structures using sequence information alone was tested by joining receptor and peptide sequences with a poly-glycine linker across two benchmark sets.
Key results
ShownAlphaFold2 with a poly-glycine linker accurately predicted bound peptide conformations from sequence alone in more than half of tested protein-peptide complexes.
Across 203 complexes from PepBDB DB and 183 complexes from PepSet, peptide C-RMSD was < 3.0 [A] in more than half of test cases and < 2.0 [A] in about 40% of cases. · Abstract
Methods, data and reproducibility
Modeled protein-peptide complex structures from sequence information alone in AlphaFold2 by concatenating receptor and peptide sequences with a poly-glycine linker and evaluating peptide C-RMSD against bound structures.
- Protein-peptide complexes from PepBDB DB · n = 203 protein-peptide complexes
- Protein-peptide complexes from PepSet · n = 183 protein-peptide complexes
- Methods & tools
- AlphaFold2, PepBDB DB, PepSet
Limitations (1)
- Nearly half of the tested protein-peptide complexes were not predicted within a peptide C-RMSD of 3.0 [A], and no head-to-head comparison against dedicated peptide-docking methods is reported in the abstract. · noted in this summary
Who should read:Read this if you model protein-peptide interactions or want a benchmark of poly-glycine linker concatenation in AlphaFold2.
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- Benchmarking Peptide Structure Prediction with AlphaFold2
A. Gulsevin, J. Meiler·bioRxiv·Feb 17, 2022
- How AlphaFold and related models predict protein-peptide complex structures
Lindsey Guan, Amy E. Keating·bioRxiv·Jun 18, 2025
- Systematic Evaluation of AlphaFold2 and OpenFold3 on Protein-Peptide Complexes
Rümeysa Fayetörbay, Ahmet Can Timuçin, Emel Timuçin·bioRxiv·Apr 24, 2026
- AF3Complex Yields Improved Structural Predictions of Protein Complexes
Jonathan D. Feldman, Jeffrey Skolnick·BioinformaticsPeer-reviewed·Jul 30, 2025
- AlphaFold encodes the principles to identify high affinity peptide binders
- Improved Peptide Docking with Privileged Knowledge Distillation using Deep Learning
Zicong Zhang, Jacob C. Verburgt, … Daisuke Kihara·ACS OmegaPeer-reviewed·Jun 16, 2025