Improve publication-to-trait links

[TDDB-limagrain]

Dear KM team,
I had one example on the Tomato KG were additional trait information could have been extracted from the literature. For example, we are looking for some genes involved in the interaction between tomato and a given fungus. We queried the KG with a protein named CEL1 and returned the following subgraph:

As you can see from its abstract header, PMID17916112 is clearly related to a susceptibility process. That could have been very nice to see the corresponding trait or phenotype or ontology connected to it. Is there any room for an improvement in your text mining to manage that? Besides, I know that text mining from scientific literature and concept extraction is a real challenge...

Thank you in advance!
Thomas