AntiFold: Structural Antibody and Nanobody Design Now Available on Vecura

What is AntiFold?

AntiFold is an advanced inverse-folding model specifically tailored for antibody and nanobody engineering. By leveraging structural coordinates from IMGT-numbered PDB files, it evaluates how well specific amino acid sequences fit a desired 3D fold. The model allows researchers to identify which residues are structurally constrained and which are tolerant to mutation, facilitating both the analysis of existing antibodies and the design of novel sequence variants.

What can users do with AntiFold on Vecura?

With AntiFold on Vecura, users can:

• Score existing sequences: Assess how well a given antibody or nanobody sequence is supported by its 3D structural context using per-residue log-probability scores.
• Generate diverse variants: Sample new sequences specifically within user-defined regions like CDR1, CDR2, or CDR3 while maintaining the global antibody fold.
• Identify mutation hotspots: Pinpoint positions that are highly constrained (and thus risky to mutate) versus those that are flexible, aiding in rational protein engineering.
• Streamline affinity maturation: Easily generate candidate sequences for experimental validation without needing to set up complex local computational pipelines.

What the output means

The output provides a detailed CSV table containing per-residue inverse-folding scores—including perplexity values and probabilities for all 20 standard amino acids—as well as FASTA files containing newly generated sequence candidates.

This output should be used to support scientific decision making. It does not replace experimental validation.

Why this matters

The ability to accurately predict sequence-structure compatibility is a cornerstone of therapeutic antibody development. Traditional methods often struggle to balance the need for structural stability with the functional necessity of diversifying antigen-binding loops (CDRs). AntiFold addresses this by providing a targeted, structure-aware approach to design.

By integrating this model into a user-friendly platform, researchers can drastically reduce the time spent on computational infrastructure and focus on the iterative process of designing antibodies with improved affinity and stability. It bridges the gap between raw structural data and actionable, laboratory-ready sequence candidates.

• Developed by: The Oxford Protein Informatics Group (OPIG)
• Source: AntiFold GitHub Repository
• Reference: AntiFold paper (Bioinformatics Advances, 2025)