AF2BIND

What is AF2BIND?

AF2BIND finds the residues on a protein most likely to contact a small-molecule ligand. It reads a single protein structure and returns a per-residue p(bind) score, so the binding pocket shows up as a cluster of high-scoring residues even when no ligand is present in the input.

The method works by borrowing AlphaFold2's learned sense of which residues sit at interfaces. Instead of docking an actual molecule, it presents the protein with a 20-residue bait made from single amino acids acting as stand-in ligands, then reads how strongly AlphaFold2's internal representation couples each target residue to those probes. Residues that the network treats as interface-like score high. This makes AF2BIND a fast way to localize a pocket before committing to docking or pocket geometry analysis.

AF2BIND scores one chain, a comma-separated set of chains, or a compact domain at a time. The selected chains can contain up to 300 residues processed by AF2BIND in total; unselected chains in the uploaded structure do not count toward that limit. Every processed residue must include its N, CA, and C backbone atoms. ProteinIQ checks this before the run because AF2BIND can otherwise write missing atoms at invalid coordinates. For a larger target, trim to the domain or interface region of interest before scoring.

This tool runs AF2BIND's standard notebook path. The experimental serial notebook and the separate large-protein domain-splitting pipeline are not included; use a trimmed domain of up to 300 residues for the standard method reviewed here.

How to use AF2BIND online

Upload a protein structure as a PDB file or fetch one from RCSB by ID, pick the chain to score, and AF2BIND returns a ranked table of p(bind) values plus a copy of the structure colored by those scores. The prediction runs on GPU infrastructure with no AlphaFold2 setup, no MSA database, and no local install. The top-ranked residues mark the most probable ligand-binding site.

The browser intentionally uses RCSB as its single database fetch source because it returns experimental PDB entries directly. API submissions may also provide an AlphaFoldDB UniProt identifier; AF2BIND then retrieves the corresponding predicted PDB structure before scoring it.

Inputs

Settings

Masking sidechains and keeping the sequence visible is the configuration AF2BIND was validated with. Turning on sequence masking is useful when testing how much a prediction depends on structure versus sequence, but it is not the recommended starting point.

Results

How AF2BIND works

AlphaFold2 builds a pair representation, a tensor that encodes the relationship between every pair of residues across all chains in a complex. When two chains form an interface, the pair representation between their residues carries a distinctive signal that the network learned from real protein structures.

AF2BIND exploits this without needing a real ligand. It adds 20 separate single-residue "bait" chains to the prediction, one for each amino acid type, using the bait sequence ACDEFGHIKLMNPQRSTVWY. Each bait acts as a minimal pseudo-ligand. AlphaFold2 then produces a pair representation linking every target residue to every bait. A small trained logistic-regression layer reads those cross-chain features and outputs p(bind), the probability that a given target residue lies in a ligand-binding pocket.

Because the baits are individual amino acids rather than a specific compound, p(bind) flags where ligands tend to bind in general, not the affinity of any particular molecule.

Interpreting p(bind)

p(bind) is a calibrated per-residue probability of belonging to a small-molecule binding site. The final AF2BIND study recommends p(bind) ≥ 0.28 as the general classification threshold; at that threshold, the reported benchmark reached 67% sensitivity, a 4.3% false-positive rate, and 63% precision. Spatial context still matters: a convincing pocket usually appears as a compact group of high-scoring residues rather than isolated values scattered across the structure.

A practical reading approach:

• Start with residues at or above 0.28, then inspect the top 10 to 15 scores and their spatial arrangement.
• Map those residues onto the displayed structure or viewer.pdb. Residues that cluster spatially are the predicted pocket; lone high scorers far from the cluster are usually noise. Use output.pdb when you specifically need the unmodified source artifact.
• Use the activation heatmap to see which amino-acid baits support or oppose each high-scoring residue; these contributions explain the model signal but do not identify a preferred real ligand.

For proteins with more than one pocket, expect more than one spatial cluster. The ranking alone will not separate them, so the structural view matters.

When to use AF2BIND vs alternatives

AF2BIND is strongest when a structure exists and the goal is to localize a small-molecule pocket quickly, especially for a protein where no bound ligand has been resolved. It needs no MSA and no docking box.

• For pockets defined by surface geometry and cavity detection rather than a learned interface signal, use fpocket, which is faster and gives explicit pocket volumes and druggability scores.
• For predicting protein-protein or protein-DNA binding residues rather than small-molecule sites, use ScanNet.
• Once a pocket is identified, estimate how a specific compound binds it with AutoDock Vina.

If the structure itself is uncertain, predict it first with AlphaFold 2, then score the model with AF2BIND.