{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://api-evangelist.com/schemas/evolutionaryscale/evolutionaryscale-esm-protein-schema.json",
  "title": "EvolutionaryScale ESMProtein",
  "description": "Schema for the `ESMProtein` object exposed by the EvolutionaryScale `esm` Python SDK and accepted by the Forge ESM3 API. An ESMProtein is the canonical multi-track representation of a protein covering sequence, structure, secondary structure, SASA, and function annotations.",
  "type": "object",
  "properties": {
    "sequence": {
      "type": "string",
      "description": "Amino acid sequence in one-letter codes (A, R, N, D, C, E, Q, G, H, I, L, K, M, F, P, S, T, W, Y, V). An underscore `_` denotes a masked position to be filled by generation.",
      "pattern": "^[ARNDCEQGHILKMFPSTWYV_X*-]+$",
      "example": "MKTAYIAKQRQISFVK_____SSERVKKLLVGDIVT"
    },
    "coordinates": {
      "type": "array",
      "description": "Per-residue atom37 coordinates, shape [L, 37, 3]. Missing atoms are represented as nulls or NaNs depending on serialization.",
      "items": {
        "type": "array",
        "minItems": 37,
        "maxItems": 37,
        "items": {
          "type": "array",
          "minItems": 3,
          "maxItems": 3,
          "items": { "type": "number" }
        }
      }
    },
    "secondary_structure": {
      "type": "string",
      "description": "DSSP-style per-residue secondary structure string. Common codes: H (alpha-helix), E (beta-strand), C (coil)."
    },
    "sasa": {
      "type": "array",
      "description": "Per-residue solvent accessible surface area (square Angstroms).",
      "items": { "type": "number", "minimum": 0 }
    },
    "function_annotations": {
      "type": "array",
      "description": "Optional list of per-residue function annotations referencing InterPro / GO / Pfam labels.",
      "items": {
        "type": "object",
        "properties": {
          "label": { "type": "string" },
          "start": { "type": "integer", "minimum": 1 },
          "end": { "type": "integer", "minimum": 1 },
          "source": { "type": "string" }
        }
      }
    },
    "plddt": {
      "type": "array",
      "description": "Per-residue pLDDT confidence score (0.0–100.0).",
      "items": { "type": "number", "minimum": 0, "maximum": 100 }
    },
    "ptm": {
      "type": "number",
      "description": "Predicted TM-score (0.0–1.0) for the structure.",
      "minimum": 0,
      "maximum": 1
    }
  },
  "anyOf": [
    { "required": ["sequence"] },
    { "required": ["coordinates"] }
  ],
  "additionalProperties": false
}