You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi,
Would DeepMod2 be able to pinpoint locations where uncharacterised modifications are present?
The current base callers can not handle this, because not trained for these.
But being able to localise where unknown modifications are present would already be very helpfull.
Thanks a ton,
Gerben
The text was updated successfully, but these errors were encountered:
Currently we have not developed any models for de-novo modification detection, but it is something we plan to do in the near future. We will update the issue once we have such models.
Currently we have not developed any models for de-novo modification detection, but it is something we plan to do in the near future. We will update the issue once we have such models.
Best, Umair
Hello, I notice that you also offer the pipline of training. So will you recommend to follow the same pipline ,just replace 5mC with the novel modification, and using fully novel modified sequence and its control to train the model for detection.
Yes, you can follow the training pipeline to train a model using modified/control datasets. You would need to specify the motif or list of coordinates to be used for training. Let me know if you run into any troubles with model training.
Hi,
Would DeepMod2 be able to pinpoint locations where uncharacterised modifications are present?
The current base callers can not handle this, because not trained for these.
But being able to localise where unknown modifications are present would already be very helpfull.
Thanks a ton,
Gerben
The text was updated successfully, but these errors were encountered: