Prodigy¶

Predicts binding affinity and related interface metrics for a protein complex. It analyzes the provided PDB structure, computes interaction features between specified chain groups, and returns a CSV table with scores such as predicted binding affinity, Kd, contact counts, NIS metrics, and temperature.

Usage¶

Use this node after you have generated or loaded a protein complex structure to estimate its binding affinity and interface characteristics. Provide the PDB(s) and specify which chains belong to which interacting molecules; the node will compute interface features and produce a consolidated score table. This is typically used after docking or complex prediction steps to prioritize candidates by predicted interaction strength.

Inputs¶

Field	Required	Type	Description	Example
pdb	True	PDB	One or more protein complex structures to analyze. Each entry represents a PDB file content (text) keyed by a name. The node computes binding metrics for each provided structure.	{"complex_1.pdb": "ATOM ... END", "complex_2.pdb": "ATOM ... END"}
chains	True	STRING	Defines which chains are considered as separate molecules (groups) for interaction analysis. Use semicolons to separate different molecules and commas to group multiple chains as a single molecule. If left empty, the whole complex is used.	A;B or A,B;C
temperature	True	INT	Temperature in Celsius used for Kd prediction.	25
distance_cutoff	True	FLOAT	Atomic distance cutoff (in Å) to define interface contacts (ICs).	5.5
acc_threshold	True	FLOAT	Accessibility threshold used for buried surface area (BSA) analysis.	0.05

Outputs¶

Field	Type	Description	Example
score.csv	CSV	CSV table summarizing predicted binding scores per input structure, including binding_affinity, kd, contacts, contact_types, nis_a, nis_c, and temperature.	design,binding_affinity,kd,contacts,contact_types,nis_a,nis_c,temperature complex_1,-8.7,3.2e-6,56,hydrophobic:34;hydrogen_bonds:12,123.4,110.2,25

Important Notes¶

Chain grouping syntax: Use semicolons to separate different molecules and commas to group multiple chains within one molecule (e.g., A;B;C or A,B;C).
Empty chains input: Leaving the chains field empty uses the entire complex without grouping.
Multiple inputs: When multiple PDBs are provided, the node evaluates each and returns a combined CSV with one or more rows.
Parameter influence: distance_cutoff and acc_threshold directly affect interface detection and BSA calculations; temperature affects Kd estimation.
Input expectations: PDB must represent a protein complex with valid chain identifiers; ensure chain labels match those referenced in the chains field.

Troubleshooting¶

No or unexpected interactions reported: Verify that chains are specified correctly (e.g., A;B vs A,B;C). Incorrect grouping can cause missing or misleading interface metrics.
Chain label not found: Ensure the chain IDs used in the chains field exactly match the labels present in the PDB file(s).
Unrealistic Kd values: Check that temperature is appropriate and that the structure represents a biologically plausible interface.
Empty output or errors: Make sure the PDB input is non-empty, properly formatted, and that at least two chain groups are defined or inferable.