-
Notifications
You must be signed in to change notification settings - Fork 21
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
treats
refactor and xDTD
#2253
Comments
Hi @dkoslicki, let's have a further discussion on this. I think it is not just as simple as choosing a different predicate, because the current xDTD training does depend on any "treat" predicates in RTX-KG2. As you know, the ground-truth data (i.e., TP and TN) we used for training is from MyChem, SemMedDB, NDF-RT, and RepoDB. They only provide pairs of drug-disease without specific predicates. So the xDTD predications are all defined as "treat" or "no treat" based on probability. |
If we need xDTD to predict predicate as well, we might need to consider what data sources we can use, such as the existing "treat" predicate in KG2? |
After discussing with David, we decide to change the predicate name from the original Here is an example:
We now have
The updated code has passed all local tests. Once it clears the CI/CD, this issue can be closed? @dkoslicki |
Here is a description of the treats refactor, which is also listed in the biolink repo here. We need to determine what, if any, changes are needed in xDTD. This could be as simple as choosing a different predicate, or as complex as inspecting answers to determine which new
treats
predicate is appropriate.The text was updated successfully, but these errors were encountered: