How to generate a consensus model for Blood Brain Barrier (BBB) permeability using OpenRiskNet services?
One of the highlights from OpenRiskNet's workshop in Amsterdam was the demonstration of how OpenRiskNet's services can be used for generating consensus models. A case in point is modelling Blood Brain Barrier (BBB) penetration. Multiple BBB models are already present on the OpenRiskNet reference site (CPSign, Jaqpot, Lazar, Weka), and an obvious question is which is best? Or maybe a better question is can I generate a consensus model from these that is better than any individual one?
This series of Jupyter Notebooks illustrates this. In particular, take a look at the notebook on the use of Dempster-Shafer theory to make consensus predictions on blood-brain barrier penetration, where two different types of consensus models are generated.
For example, here are some predictions using the Yager consensus model:
For more info take a look at the notebooks mentioned above.