Uncovering Alternative Protein Folds: AF-Cluster is Now Available on Vecura

What is AF-Cluster?

AF-Cluster is a specialized protocol designed to uncover "fold-switching" proteins—proteins capable of adopting two or more structurally distinct native conformations. Traditional protein structure prediction tools, such as standard AlphaFold2 runs, often collapse all evolutionary data into a single, dominant conformation. AF-Cluster overcomes this by partitioning the multiple-sequence alignment (MSA) into evolutionary subfamilies using DBSCAN clustering, allowing researchers to predict distinct folds hidden within the evolutionary record.

What can users do with AF-Cluster on Vecura?

With AF-Cluster on Vecura, users can:

• Partition MSAs: Automatically split large, complex MSAs into coherent evolutionary sub-clusters to isolate distinct conformational signals.
• Probe Structural Diversity: Run structure prediction pipelines on individual clusters to reveal alternative protein folds that were previously suppressed.
• Analyze Ensembles: Use the built-in ensemble analysis tools to aggregate predicted PDB structures, identify distinct conformational states, and extract representative models.
• Predict Contacts: Perform independent co-evolutionary analysis using ESM-MSA-1b to validate structural contacts and support findings from the AlphaFold2-based pipeline.

What the output means

The output provides a set of clustered MSA files, comprehensive metadata regarding evolutionary sub-families, and a structural ensemble analysis. This includes identified conformational states, representative PDB structures, and statistical confidence scores.

This output should be used to support scientific decision-making regarding protein structure and function. It is a computational tool intended to guide research and does not replace experimental validation, such as X-ray crystallography or NMR spectroscopy.

Why this matters

The discovery of fold-switching proteins has significant implications for understanding biological signaling, gene regulation, and disease mechanisms, as these proteins often act as molecular switches that regulate critical cellular processes. By systematically identifying these states, researchers can gain deeper insights into how structural plasticity drives function.

AF-Cluster bridges the gap between massive sequence databases and the nuanced understanding of protein folding, providing a streamlined pathway to investigate these dynamic and biologically essential structural shifts.

• Developed by: H. Wayment-Steele et al.
• Source: Nature 625, 832–839 (2024)
• Reference: Official GitHub Repository