Closed Moeinh77 closed 1 year ago
Nope you did not miss it. We have no examples on this, yet, sorry. But there are some lines of work that might be useful as inspiration, e.g., D-Script or Topsy-Turvy. But I think when you want to approach this topic, the most important part is actually not the machine learning part but the dataset preparation part, so maybe take a look at this: More challenges for ML in PPI prediction
Thank you very much for sharing those links. My thoughts were to use the last hidden state of the networks as generated sequence features and based on the attained features for the two sequences (maybe combining them or doing another operation on 2 vectors), I do a binary classification for PPI prediction.
Perfectly valid approach and def. very interesting! - If you get somewhere with this approach, it would be great if you could share your insights here :)
How can I perform protein-protein interaction using the models? are there any examples in the repo that I couldn't find?