Boltz Predict

Runs a full Boltz structure prediction job using a prepared Boltz YAML configuration and its auxiliary files. It submits the job to the Salt biotech backend and returns predicted structures along with confidence metrics, and optionally affinity metrics when requested in the YAML.

Usage

Use after assembling your sequences, constraints, templates, and optional properties into a Boltz YAML bundle (typically via Boltz YAML Combiner). Provide that YAML and its auxiliary files to this node, set the random seed and optional inference parameters, then run to obtain predicted PDBs and confidence JSON. If your YAML includes affinity properties and ligand entities, an affinity JSON will also be returned.

Inputs

Field	Required	Type	Description	Example
boltz_yaml	True	BOLTZ_YAML	Structured configuration describing sequences (proteins/DNA/RNA/ligands), constraints, templates, and optional properties for the Boltz run. Typically produced by Boltz YAML Combiner.	{'version': 1, 'sequences': [{'protein': {'id': 'A', 'sequence': 'ACDE...'}}]}
boltz_files	True	BOLTZ_FILES	Auxiliary files referenced by the YAML (e.g., MSA files, template PDB/CIF files) packaged as a filename-to-content mapping.	{'msa_1.a3m': '', 'template_1.pdb': ''}
seed	True	INT	Base random seed used for the prediction process.	42
recycling_steps	False	INT	Number of recycling iterations during inference.	3
sampling_steps	False	INT	Number of diffusion sampling steps.	200
diffusion_samples	False	INT	How many independent diffusion samples to generate.	1
max_parallel_samples	False	INT	Maximum number of samples processed in parallel.	5
step_scale	False	FLOAT	Diffusion step scale (temperature-like parameter).	1.638
output_format	False	CHOICE	Structure output format.	pdb
num_workers	False	INT	Worker processes (0 disables multiprocessing).	0
max_msa_seqs	False	INT	Maximum number of MSA sequences to use.	8192
subsample_msa	False	BOOLEAN	Whether to subsample MSA sequences.	False
num_subsampled_msa	False	INT	Number of MSA sequences to keep when subsampling is enabled.	1024
use_potentials	False	BOOLEAN	Enable inference-time potentials for improved quality (may be slower).	False
write_full_pae	False	BOOLEAN	Write the full Predicted Aligned Error (PAE) matrix.	False
write_full_pde	False	BOOLEAN	Write the full Predicted Distance Error (PDE) matrix.	False
affinity_mw_correction	False	BOOLEAN	Apply molecular weight correction for affinity output (only used if affinity is predicted).	False
sampling_steps_affinity	False	INT	Number of sampling steps for affinity prediction.	200
diffusion_samples_affinity	False	INT	Number of diffusion samples for affinity prediction.	5

Outputs

Field	Type	Description	Example
structures.pdb	PDB	Dictionary mapping generated structure names to PDB (or chosen format) contents for ranked predictions.	{'ranked_0.pdb': '', 'ranked_1.pdb': ''}
confidence.json	JSON	Confidence metrics (e.g., per-model scores, optional matrices when enabled).	{'ranking_confidence': {'ranked_0': 0.78}, 'pae': ''}
affinity.json	JSON	Affinity prediction outputs when affinity is requested in the YAML and ligands are present; empty otherwise.	{'predicted_affinity': {'complex_0': {'kd': 1.2}}}

Important Notes

• Affinity requirements: Affinity outputs are only produced when the YAML includes affinity properties and at least one ligand sequence; otherwise affinity.json will be empty.
• Validation: The node validates that boltz_yaml is a dictionary including at least one sequence and that boltz_files is a dictionary; invalid inputs raise errors.
• Performance settings: Increasing sampling_steps, diffusion_samples, or enabling use_potentials can improve quality at the cost of runtime.
• Output format: output_format controls structure serialization (pdb or mmcif).
• MSA controls: Use max_msa_seqs, subsample_msa, and num_subsampled_msa to manage memory/runtime for large MSAs.
• No trajectory output here: This node does not return diffusion trajectory data; use Boltz Partial Diffusion for trajectories.

Troubleshooting

• Error: 'Boltz YAML must be a dictionary': Ensure you pass the YAML object produced by Boltz YAML Combiner, not a string or malformed data.
• Error: 'Boltz YAML must contain at least one sequence': Add at least one protein/DNA/RNA/ligand sequence in the YAML.
• Error: 'Boltz files must be a dictionary': Provide auxiliary files as a mapping of filenames to file contents.
• Error: 'Affinity prediction requires at least one ligand sequence': Add a ligand entity to sequences and include appropriate affinity properties in the YAML.
• Timeouts or long runs: Reduce sampling_steps or diffusion_samples, disable use_potentials, lower max_msa_seqs, or enable subsample_msa.
• Empty outputs: Check that the YAML references auxiliary files by the expected filenames and that boltz_files contains those filenames with valid contents.