HPO text-mining terms do not identify "Not" or "No" for terms
LCCarmody opened this issue · comments
When using the text mining tool, there is no identification of 'not'/ 'no' phenotypes.
Below is an example of what I added to the text-mining box. It does not matter if I put 'no' or 'not' in front of the term. It identifies them always as a present phenotype and as far as I know, from that box, I cannot further negate it.
Psychomotor retardation
Developmental regression (from age)
Seizures/ epilepsy
Lennox Gestaut
Myopathy
Abnormal basal ganglia
Cerebral atrophy
Elevated serum lactate
elevated lactate, elevated malate, elavated succinate
NOt Ataxia
Not Dystonia
Not Pyramidal tract involvement
Not Hepatomegaly
Not Cardiomyopathy
Not Leukodystrophy
I think that the reason for not identifying negative terms is that we are using another text mining service than we were using before. This service apparently does not identify the negated terms.
However, from some time ago, it is also possible to drag & drop the term check box between YES and NOT sections. You still need to check the box if you want to add it to the curated case.
Docs of this app are still somehow missing and we failed to let you know about these changes. I am sorry for that :{
Does this resolve issue?
Ah! Okay. I wasn't aware of that feature. It appears to work. Thanks