Third step: Use Textpresso-like resource to identify papers that could provide evidence for statements that aren't supported by evidence from existing annotations.

[ukemi]

When a model is created and evidence from existing annotations cannot be used as evidence for the model, it will be helpful to curators if we can mine the existing literature corpus for papers or statements that will provide evidence.

[vanaukenk]

1. If evidence for an assertion about a gene does not exist in standard annotations, we want to have a facile way for curators to search the literature for their organism to see if experimental evidence exists.
2. To do this with Textpresso, they'll need: a) a Textpresso instance for their organism, e.g. mouse, with the full text of their literature corpus (can start with the PMC Open Access subset that the Alliance already has, but will want to have the full corpus); b) need lists of relevant mouse entities or vocabularies.
3. If a match can't be found in the organism-specific Textpresso instance, then having a Textpresso on the entire PMC Open Access subset would be a next step for searching.