Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[bioontology-support] SNOMED circular parent

Ted Bashor ted at wonderlakesoftware.com
Mon Feb 24 11:45:51 PST 2014


I’ve confirmed that the strange parent class relationships I was seeing last week have gone away, and that I’m seeing classes that were added in the newer version of SNOMEDCT.
Thanks! Ted

On Feb 24, 2014, at 10:14 AM, Paul R Alexander <palexander at stanford.edu> wrote:

> Ted,
> 
> This issue should be corrected. The 2013AB distribution of UMLS contained a combined US/International version of SNOMEDCT that added some relations we weren’t handling properly. I believe we’ve corrected this problem and we’ll be looking into other things as they arise.
> 
> Thanks for reporting the issue.
> 
> Paul
> 
> 
> On Feb 20, 2014, at 9:14 AM, Ted Bashor <ted at wonderlakesoftware.com> wrote:
> 
>> Thanks for the update.  Looks like it isn’t just one or two cases — it’s bad in a lot of places like the one below.  Too many to work around.  Any chance you can revert to the previous SNOMED index while this is being sorted out?
>> 
>> Disorder of upper digestive tract : 50410009
>>   Disorder of digestive tract : 84410009
>>     Disorder of digestive organ : 76712006
>>       Disorder of digestive tract : 84410009
>>         Disorder of digestive organ : 76712006
>>           ...
>> 
>> The version says it is 2013AA.  Would be great if the more recent UMLS 2013AB release were available.   (There’s actually a 2014 SNOMED release, but sounds like you only use the UMLS distributions)
>> 
>> Thanks, Ted
>> 
>> On Feb 19, 2014, at 5:13 PM, Ray Fergerson <ray.fergerson at stanford.edu> wrote:
>> 
>>> Yes, this does appear to be wrong. The best that I can tell is that “Disorder of connective tissue” really has a single parent “Disorder by body site”.  I don’t know why this loop with “Musculoskeletal ….” is getting created.
>>>  
>>> Unfortunately the expert on these sorts of bizarre problems just went on vacation. It will probably be late next week before this gets sorted out. Sorry.
>>>  
>>> Ray
>>>  
>>> From: bioontology-support-bounces at lists.stanford.edu [mailto:bioontology-support-bounces at lists.stanford.edu] On Behalf Of Ted Bashor
>>> Sent: Wednesday, February 19, 2014 4:04 PM
>>> To: support at bioontology.org
>>> Subject: [bioontology-support] SNOMED circular parent
>>>  
>>> Looks like the version of SNOMED was updated yesterday.  I’ve run into a few funky things with new inferred relationships, but the one that is most serious is a circular parent/subClassOf.  Walking up the parent hierarchy from Soft tissue lesion, you find "Musculoskeletal and connective tissue disorder” has a parent of "Disorder of connective tissue” which has a parent "Musculoskeletal and connective tissue disorder” etc.
>>>  
>>> Soft tissue lesion : 239953001
>>>   Rheumatism : 29913006
>>>     Arthropathies and related disorders : 201433006
>>>       Musculoskeletal and connective tissue disorder : 312225001
>>>         Disorder of connective tissue : 105969002
>>>           Musculoskeletal and connective tissue disorder : 312225001
>>>             Disorder of connective tissue : 105969002
>>>               ...
>>>  
>>> I haven’t dug into the UMLS source for this release yet, any idea if this could be a bug in bioportal’s indexing.
>>> Thanks, Ted
>> 
>> _______________________________________________
>> bioontology-support mailing list
>> bioontology-support at lists.stanford.edu
>> https://mailman.stanford.edu/mailman/listinfo/bioontology-support
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/bioontology-support/attachments/20140224/c5208b4d/attachment-0001.html>


More information about the bioontology-support mailing list