starplast.iedb

Antibody epitopes per gene, from IEDB directly rather than through ToxoDB.

ToxoDB integrates IEDB and exposes a count per gene, which is what fills T-cell epitope content. That count is not split by epitope type. IEDB's own API is, and the split is the whole point here: seroreactivity / antigenicity is a question about ANTIBODIES, and answering it with a number dominated by T-cell assays would be answering a different question with a plausible column.

bcell_search is the antibody half. Filtering it to Toxoplasma source antigens gives 939 assay records over 44 antigens, which reduce to 241 distinct epitope sequences.

Distinct epitopes, not assay records

A protein studied by twenty groups accumulates twenty records for the same peptide, so counting records would rank antigens by how fashionable they are. Counting distinct sequences asks how much of the protein antibodies actually recognise, which is what the slot means by antigenicity.

Why the mapping is by product description

IEDB identifies antigens by UniProt accession and a name, and the names are verbatim ToxoDB product descriptions -- SAG-related sequence SRS29B, Dense granule protein GRA6. Matching those exactly resolves 34 of 44. Matching by the trailing symbol instead resolves only 24 and loses SRS29B, which is SAG1 and the single most studied antigen in the organism, so the description is the better key and the symbol is only a fallback. The eleven that remain are generic -- Uncharacterized protein, PA14 domain-containing protein -- and match many genes each, so they are dropped rather than guessed.