{
  "sources": [
    "https://ftp.ebi.ac.uk/pub/contrib/UniProt/ProtNLM2/List_of_UniProt_accessions_that_have_ProtNLM2_annotations.tsv",
    "https://rest.uniprot.org/uniprotkb/protnlm/{accession}",
    "https://rest.uniprot.org/uniprotkb/{accession}.json"
  ],
  "checksums": {
    "selection.csv": "66d8a18bf6345d61e7556eac6a1cfc2e7749f7cf253c68495dd923a93daa4eda",
    "api-snapshot.jsonl.gz": "9449ba689c38f40203ecac5df4049a84841acc2f15d026baf882bca40f921fcc",
    "reference-searches.jsonl.gz": "d639493c7c7fb5b8e1d767a6e6f628d3c7285fc60a94dcc15faa9678078bb7bc",
    "published-accessions.tsv.gz": "1394f39fe0f0bd6ee93965a233b918f45dc03f3e125ab5e94e5a2cbe1152edb6",
    "cohort.csv": "cd406b901dd83ffdb821d6cf0d01888a08d17f3c34a1b203f33dc475abb0d81e",
    "prediction-statements.csv": "138125202b11ab614763db2938389322f2d610dcbc5f2048071c17b5c892904b",
    "same-gene-references.csv": "8765cb4bbbd1a45e5d9bae36cff9299f26283f1c9bd09d857a94f8b9c528afe4",
    "sequences.fasta": "bebf27529a612546612ed706440938278d03a7c5fb6610cfc54287c6ce500224",
    "summary.json": "6ae86bce938caa76faa58f378947c0df33b01b88e29273322437dc08743688fe",
    "summarize.py": "6de619177adeae03381eb899ca8c38c5fa1c8ef355e9f520a9a705c56df1517d"
  },
  "selection": "Purposive cross-species challenge set selected after inspecting predictions; not a random sample or an estimate of accuracy.",
  "sequence_scope": "Current exact-accession UniProt sequences. Prediction responses do not include the original model input sequence; identity to that historical input is not established.",
  "reference_scope": "Reviewed same-gene candidates require a shared MOD identifier. They are comparison records, not substitute prediction targets or automatic validation.",
  "review_scope": "Selection only. Proposed evolutionary comparators are not established orthologs or completed analyses."
}