corinabioinformatic
commented
3 weeks ago Same issue here. The impact is getting huge taking account that some websites, when mapping to a specific term, can miss a huge amount of data.
e.g: I am interested in "Invasive breast carcinoma".
1) I go to Oncotree and look for it: https://[oncotree.mskcc.org/#/home](https://oncotree.mskcc.org/#/home)
Great. it exists.
2) I want to get information from the "ProjectScore" from cell lines related to this disease. so i go to opentargets and I do a search by the disease term "Invasive breast carcinoma: https://platform.opentargets.org/disease/EFO_1000307/associations
3) Surprisingly there is no information related. so i go directly the the project source and type my fav cell line related to "invasive breast carcinoma"
And IT DOES NOT FIND IT!
But if i just type: Breast Carcinoma
it find it!
4) but, is my favorite cell line "breast carcinoma"? or "invasive breast carcinoma"? So then i will pick my favorite cell line (that i though it was "invasive cell carcinoma") --> SUM-52PE.
I see the oncology term is "Breast carcinoma"
5) I am aware this Depmap source is somehow related to Depmap "Broad institute". So i go and check for this cell line in there too. Surprisingly, my cell line appears with a different oncotree term. this time, is correctly mapped to the term I found originally in Oncotree website.
CONCLUSION:
1) Is oncotree website/mapper missing some parent nodes? (NOTE: "Invasive breast carcinoma (BRCA)" and "Breast Carcinoma" both appears in the NCIT (national cancer institute thesaurus) : breast cancer: C4872 and invasive breast cancer- C9245 . THis will lead to missing mappers from tools like opentargets or sangerDepmap.
2) On the other hand, is the problem from the SangerDepmap, who are using maybe outdated annotations, because Oncotree is not considering "breast carcinoma" anymore? (this ticket will be shared with the responsibles of this project Project Score. so maybe you both can reach a consensus and please keep us in the loop).
3) Also I will contact Opentargets team because maybe I am missing any important information from their documentation? Although in the community not question has been raised about oncotree so far.
Please keep us in the loop here.
Many thanks!
Following up on an email I sent to the mailing list about mappings to other resources like NCIT, UMLS and DOID, I noticed that the mappings file at https://github.com/cBioPortal/oncotree/blob/master/scripts/ontology_to_ontology_mapping_tool/ontology_mappings.txt provides mappings for 771 non-tissue OncoTree terms to NCIT while the data returned from the web service at https://oncotree.mskcc.org:443/api/tumorTypes/tree only provides mappings for 525 non-tissue OncoTree terms to NCIT.
When I first looked at the data from the webservice I concluded that a lot of mappings are missing but the mappings file on GitHub appears to be much closer to complete. Is there a reason why not all mappings from the file are made available in the web service?