"damage" matched as entity about processors HOT 5 CLOSED

I've prepared an exhaustive list of words that overlap between capitalized protein names and English words. This isn't filtered for word frequency, so use with care -- aah is unlikely to appear in a biomedical text except as a protein name, but that might not be the case for cat, ape, chimp, etc.
overlap.txt

from processors.

Hi @danebell: I would like to add this list to bioresources and use it to filter out false positives in the rule-based NER. But, at this point, it seems too over-inclusive. Can you please do another pass over it, and keep just words that are likely to appear as non-proteins in bio texts? For example, I would remove the first 9 entries (they are likely to be only used as proteins), but keep ACT, ANT, etc.

from processors.

Yes, I can definitely do this, but there are over 1000 entries, so if it's okay, I will wait for at least a week to do so.

from processors.

Agreed, it can wait. I suspect that less than 10% of this are actually
relevant in this list (the rest are valid protein names that are unlikely
to be used otherwise in bio texts).

On Wed, Feb 24, 2016 at 11:03 AM, Dane Bell [email protected]
wrote:

Yes, I can definitely do this, but there are over 1000 entries, so if it's
okay, I will wait for at least a week to do so.

—
Reply to this email directly or view it on GitHub
#25 (comment).

from processors.

closed by the addition of the NER stop list in bioresources 1.1.9.

from processors.