Unlock Protein Modeling with Noncanonical Amino Acids via RareFold on Vecura

What is RareFold?

RareFold is an advanced computational framework designed to predict the 3D structures of single-chain proteins containing noncanonical amino acids (NCAAs) and to design peptide binders incorporating these residues. By extending traditional structure-prediction architectures, it processes a heterogeneous alphabet of 20 canonical and 29 rare amino acid types. It helps users model chemically modified proteins and discover novel therapeutic peptide binders with enhanced structural properties. It is especially useful for researchers studying post-translational modifications, selenoproteins, or designing custom peptide-based therapeutics that require specific chemical functionality.

What can users do with RareFold on Vecura?

With RareFold on Vecura, users can:

• Predict the 3D structure of single-chain proteins containing one or more NCAAs from a provided sequence.

• Design linear or cyclic peptide binders against a target protein sequence using the EvoBindRare Monte-Carlo optimization loop.

• Fine-tune design parameters, such as binder length, specific NCAA availability, and cycle settings to explore chemical space.

• Analyze structural confidence via per-residue pLDDT scores and optimize binding interfaces based on detailed metrics like interface distance and steric clash fractions.

What the output means

The output provides predicted 3D structures in PDB format, accompanied by per-residue pLDDT confidence scores. For binder design, users receive a comprehensive set of design metrics, including binding trajectories and optimized complexes.

This output should be used to support scientific decision making. It does not replace experimental validation, and users should treat these computational results as a starting point for subsequent in-vitro or in-vivo testing.

Why this matters

The inclusion of noncanonical amino acids—such as phosphorylated, methylated, or selenized residues—is critical for understanding biological signaling and creating next-generation therapeutics. Traditional protein folding models often ignore these modifications, limiting their utility in complex biological scenarios.

RareFold bridges this gap, allowing researchers to incorporate the diverse chemistry of NCAAs into structural modeling and design. This capability facilitates the development of sophisticated peptide binders that can mimic natural interactions or achieve chemical stability unattainable with only standard amino acids.

• Developed by: Patrick Bryant and collaborators

• Source: Official GitHub repository

• Reference: RareFold paper (bioRxiv 2025)