Discovery / research surface
Search the graph.
One entry point for the whole substrate. Type a gene (human or dog), a human variant, a Dog10K position, a breed, or a human disease. A gene or variant resolves right here against the Dog10K → human disease map; a breed or a human disease opens its lens. When the answer is nothing, it always tells you why.
Human → dog is first-class. Ask "does any Dog10K dog carry a variant
at this human disease position?" and you get the orthologous-position answer. We report where a variant
sits, never that a dog allele is identical to a human one (that is a separate, harder claim, INV-57).
Want the ranked landscape instead of a single lookup? See the full candidate list →
How "not found" answers
Every empty result returns exactly one reason, the first rung that failed, in order:
- 1Not in the disease-gene set, no human P/LP ClinVar evidence to bridge from. (not "the dog is clear")
- 2No one2one ortholog, a human disease gene, but the dog ortholog is unmapped or one-to-many; the bridge requires one2one (INV-56).
- 3Nothing cleared the stack, anchored, but no Dog10K variant in the gene passed the evidence gate. (gene-level; the per-variant rung, bare position vs mechanism mismatch, is the next data drop)
- ✓Cleared, a Dog10K variant sits at a human disease position and cleared the stack. A candidate, never a confirmed model.