I have a huge list of symptoms (1 million rows+ ). I would like to reduce the data by grouping the symptoms. For example, lower abdominal pain and upper abdominal pain into "abdominal pain" category. Is there any free API or downloadable database available to refer the hierarchy of diseases and symptoms for grouping? Many Thanks

  • This does not answer your question about hierarchy, but a tip for matching terms like lower abdominal pain and upper abdominal pain into one term is OpenRefine
    – user4293
    Feb 17 '17 at 15:04

The NIH has published the UMLS database, which consists of more than 7 million concepts, diseases and symptoms. It's a very wealthy resource. The license is pretty permissive if you are working in the United States. Check it out on the UMLS website. It's completely free and curated by the National Library of Medicine.


I am not exactly sure that this is what you are referring to, but the MEDDRA database costs money in most cases, but I believe there are some research licenses. It is a hierarchy of indications.


Hope that helps.

  • Thanks Hans for your quick reply. Yes. I am aware of MEDDRA . I am looking if there are any free sources. Thanks for you suggestion though.
    – Breeze S
    Jul 8 '15 at 13:59
  • Yeah, I don't think that there is a free one. If you find one, post back here though! :-) Jul 8 '15 at 22:33
  • @BreezeS That'd be great if you could indicate in your question which data sets you're aware of. Jan 8 '16 at 22:32

Have a look at the supplementary material of Learning disease relationships from clinical drug trials. disease_mappings10.txt file in the supplementary contains the info you need.

For instance, if you do a search (e.g., grep "lower abdominal pain" disease_mappings10.txt), you will see

lower abdominal pain abdominal pain Approximate


Check Bioportal for an ontology suiting your needs. There are various disease, anatomy and symptom related ontologies out there, which might help you there.

Use the European Bioinformatics Institute Ontology Lookup Service (EBI OLS), to get suggestions, which ontologies you could use.

  • Welcome to stackexchange. These are good suggestions. If you want to improve your answer further you might want to consider providing direct links to the Bioportoal and the EBI OLS in your post. Feb 15 '17 at 7:23
  • You can also use obofoundry.org - the set is smaller than bioportal, but all should be open and in principle the goal is for the ontologies to work together. For your needs (symptoms), I would look at HP first (phenotypes) and for diseases/disorders DOID and MONDO (all these also visible on OLS and BioPortal)
    – cmungall
    Jan 8 '19 at 3:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.