search

WormBase / ACKnowledge

Author Curation to Knowledgebases
MIT License
1 stars 1 forks source link

'C. tropicalis' not recognized with simple thresholds and tfidf #165

Closed draciti closed 2 years ago

draciti commented 4 years ago

for WBPaper00054685. @valearna , when you get a chance can you please take a look?

valearna commented 4 years ago

Sure thing @draciti. It looks like a text conversion issue. I can't find any mentions to 'tropicalis' in the converted txt.

draciti commented 4 years ago

Thanks @valearna, that's what i suspected but wanted to make sure. I don't think we can do much for now. Closing. We will revisit when we will talk about taking text from Open Access. thx!

draciti commented 2 years ago

@valearna I am reopening this issue, can we look a little closer on why tropicalis was not picked up. Not sure if there's much to do but PubTator and EuropePMC extract it. thanks

valearna commented 2 years ago

@draciti I can re-process this paper with the new pdf2txt library that we are now using in production and see if tropicalis gets picked up. I'll run it on mangolassi.

valearna commented 2 years ago

Can you take a look at the species found by using the same algorithm that is now in prod? Here's the link: https://tinyurl.com/yge9x9ke Note that the author already submitted data for this paper. I have deleted the list of species submitted by the author on mangolassi to show you the results of the extraction directly in the form

vanaukenk commented 2 years ago

The extraction now picks up Caenorhabditis tropicalis but also Candida tropicalis. I don't know if we've encountered a similar situation before, i.e. two possible species matches for the same genus and species abbreviation, e.g. 'C. tropicalis'.
Only the Caenorhabditis tropicalis is correct..

valearna commented 2 years ago

We had a similar case with Caenorhabditis elegans and Cunninghamella elegans and we decided to blacklist Cunninghamella elegans. We could do the same with Candida tropicalis.

vanaukenk commented 2 years ago

@valearna I think it would be okay to also blacklist Candida tropicalis. At least for now. @draciti what do you think?

draciti commented 2 years ago

Agreed, we can blacklist Candida tropicalis. Thanks Valerio for looking into this. @valearna , once you add Candida tropicalis in the blacklist, you can close this ticket

valearna commented 2 years ago

Added taxon ID 5482 to exclusion list

valearna commented 2 years ago

Thanks @draciti and @vanaukenk. I'll test the new pipeline on mangolassi before closing

valearna commented 2 years ago

Tested - it works. Closing