The goal of this project is to predict possible reference biological pathways, or extend current ones, using machine learning.
First, we classify experimental interactions as functional or not, to decide if they are good candidates to be part of a biological pathway.
Second, we take sets of consecutive interactions and decide if they resemble a pathway, or extend a current one.
The project is open source, free to use and extend. Contributions are welcome.
More details can be found in the Wiki.