A consensus database of the accessible human surfaceome.
Five public surfaceome databases disagree more often than they agree. This atlas reconciles them: an LLM agent triages every protein-coding human gene, and a deep-dive agent compiles cited evidence for each strong candidate.
Curated shortlists over the deep-dive cohort. Canonical is the strictest tier, Likely is broader, Cell-state induced and Cell-type restricted are sub-buckets of Likely.
Applies only to genes with a deep-dive record. Non-deep-dive rows auto-exclude because the predicate reads fields the catalog row doesn't carry. The count badge on each chip is the population that survives the predicate alone — refine further with the existing search and facet chips below.
The high-confidence surface shortlist: a confident overall call — surface-dominant or mixed, accessible, dense evidence (at least supportive-but-indirect; weak and conflicting excluded). Full gate on the API page.
Applies only to genes with a deep-dive record. Non-deep-dive rows auto-exclude because the predicate reads fields the catalog row doesn't carry.
Broader surface set than Canonical — meets the same surface-accessibility bar but falls short on other criteria: admits predominantly intracellular proteins with a surface fraction (e.g. SRC via lysosomal exocytosis, HMGB1 via DAMP release) and stronger state-dependence. This chip excludes Canonical genes; every Canonical gene is also Likely, so select both chips to see the full Likely tier.
Applies only to genes with a deep-dive record. Non-deep-dive rows auto-exclude because the predicate reads fields the catalog row doesn't carry.
Subset of Likely that reaches the surface because of a cell-state change — activation, stress, or oncogenic transformation — rather than constitutively (SRC, CD63, HMGB1, C3). Defined by the deep dive's surface-mechanism call (genuine state-induced surfacing), not by mere tumour association; the oncogenic / immune / stress / infection trigger is shown as the sub-chips.
Applies only to genes with a deep-dive record. Non-deep-dive rows auto-exclude because the predicate reads fields the catalog row doesn't carry.
See the exact Cell-state induced gate definition on the API page →
Subset of Likely with constitutive surface in specific cell types only (KLK2 in prostate, etc.). Different cell types — not same cell across states.
Applies only to genes with a deep-dive record. Non-deep-dive rows auto-exclude because the predicate reads fields the catalog row doesn't carry.
See the exact Cell-type restricted gate definition on the API page →
Non-canonical genes with a thin discovery corpus (< 100 papers found) that UniProt nonetheless annotates as cell-surface. These are under-studied surface candidates — the deep dive likely couldn't reach a confident call for lack of literature, not because the protein is intracellular. Good re-dive targets. (UniProt is the surface database used here because it outperformed the others on our gold-standard positive controls.)
Applies only to genes with a deep-dive record. Non-deep-dive rows auto-exclude because the predicate reads fields the catalog row doesn't carry.
yes verdict can only carry a yes-bucket reason. This is also the reason the filter chips above filter on.surface_call_reason), re-derived from the full evidence ledger rather than abstracts. It is bound to the deep-dive verdict and state-dependence, not to the triage verdict, so it legitimately disagrees with triage for roughly a third of deep-dived genes — that disagreement is the deep dive doing its job, not a data error. Blank where no deep dive has run.