Database

​

Molecule

• Molecules are composed of chains, which are protein sequences, small molecules, or nucleic acids.
• A molecule is a unique set of chains. A molecule can be part of multiple groups, and its tool scores accumulate across all of them.
• Molecules can have associated structure files, such as a crystal structure or a structure predicted by a tool, each labeled by its source.
• Molecules are unique within your organization, but you may re-upload one as many times as you like. If you re-upload the same set of chains, the same underlying molecule is used and its associated scores are kept.

// GET https://app.tamarind.bio/api/molecules
{
  "items": [
    {
      "id": "3f8a1c2e-5b7d-4e9f-a1b2-c3d4e5f6a7b8",
      "name": "anti-PD-L1 binder",
      "type": "protein",
      "entity": { "H": "EVQ...", "L": "DIQ..." },
      "chainMapping": {
        "H": { "type": "protein", "tags": ["heavy"] },
        "L": { "type": "protein", "tags": ["light"] }
      },
      "files": {
        "pdb": [
          { "fileName": "model.pdb", "fileType": "pdb",
            "downloadUrl": "https://...", "createdAt": "2026-01-02T00:00:00Z" }
        ]
      },
      "metadata": {
        "alphafold": [{ "jobId": "7efa949f-...", "ptm": 0.86, "plDDT": 88.1 }],
        "affinity_nM": 12.4
      },
      "origin": { "type": "job", "jobId": "7efa949f-...", "jobType": "alphafold" },
      "source": null,
      "groups": [{ "id": "8b1d...", "name": "PD-L1 binders" }],
      "truncated": []
    }
  ],
  "nextCursor": null
}

​

Importing molecules

• Tamarind tools which generate new molecules — protein design tools, database search tools, or small molecule design tools — automatically save their outputs in the Tamarind Molecular Database.
• You may import molecules from CSV, FASTA, PDB, SMILES, plain text, or by drawing one, at /molecules/new.
• A molecule’s sequences may not be edited after creation.
• A molecule’s source is the molecule it was created from — for example, the original molecule before affinity optimization, or before making point mutations. A de novo designed binder has no source, as does a user-uploaded small molecule.
• A molecule’s origin is where it was created — for example, a job/pipeline or a user.

​

Molecule Groups

• A group is a collection of molecules.
• Tools which generate new designed molecules — diversification tools (e.g. ProteinMPNN) or de novo design tools (e.g. RFantibody) — create new output molecule groups for the job. Structure prediction and scoring tools attach their results to existing input molecule groups (one is created automatically if not already present).
• Chain IDs and their roles (e.g. heavy/light for antibodies, or lead to distinguish your primary drug from targets/cofactors) are saved at the group level. This means a molecule can have different chain IDs across different groups.
• Optionally, you may bind schemas to groups to enforce data types and required chains/fields.