Protenix V1 Tools#

Configuration File: remote_tools/protenix_v1_tools.json Tool Type: Remote Tools Count: 1

This page contains all tools defined in the protenix_v1_tools.json configuration file.

Available Tools#

protenix_v1_fold (Type: RemoteTool)#

Predict an all-atom biomolecular structure with Protenix 1.1.0, ByteDance’s open AlphaFold3 repro…

protenix_v1_fold tool specification

Tool Information:

  • Name: protenix_v1_fold

  • Type: RemoteTool

  • Description: Predict an all-atom biomolecular structure with Protenix 1.1.0, ByteDance’s open AlphaFold3 reproduction. The query is a list of chains, each with a molecule_type (protein, dna, rna, ligand or ion) and a sequence – SMILES or CCD_<id> for a ligand. Returns per-sample confidence scores (rank on ranking_score) and timings; the mmCIF text itself is opt-in. No MSA search is performed: pass alignments on a chain for an MSA-fed prediction.

Parameters:

  • chains (array) (required) The chains of one complex.

  • name (string) (optional) A label for this complex; it names the output files. Defaults to ‘complex’.

  • seed (integer) (optional) Random seed; the same seed reproduces the prediction. Upstream’s default is 101.

  • num_diffusion_samples (integer) (optional) Structures to sample (upstream’s default is 5). Each is scored separately, so this is how to compare conformations; rank them on scores.

  • num_diffusion_steps (integer) (optional) Diffusion steps per sample (upstream’s default is 200). Fewer is faster and coarser.

  • num_recycles (integer) (optional) Pairformer recycling cycles (upstream’s default is 10). Fewer is faster and less accurate.

  • out_dir (string) (optional) Write the predicted structures to this directory on the server instead of a temporary one, and return their paths. This is how to get a structure larger than the transport can carry, and how to hand one provider’s output to another’s input without the mmCIF making a round trip through the caller.

  • return_structures (boolean) (optional) Return the mmCIF text (default false: megabytes). The confidence scores come back either way.

Example Usage:

query = {
    "name": "protenix_v1_fold",
    "arguments": {
        "chains": ["item1", "item2"]
    }
}
result = tu.run(query)