Thank you for your outstanding work on CLEAN. In my current research, I am attempting to design enzyme proteins from scratch and using CLEAN to predict EC numbers. At what confidence threshold would you recommend further experimental validation of the predicted sequences? Additionally, I have noticed that in my design process, when CLEAN predicts an EC number with a confidence score greater than 0.8, the corresponding sequences often have low final pLDDT scores when predicted using AF3.
Dear Author,
Thank you for your outstanding work on CLEAN. In my current research, I am attempting to design enzyme proteins from scratch and using CLEAN to predict EC numbers. At what confidence threshold would you recommend further experimental validation of the predicted sequences? Additionally, I have noticed that in my design process, when CLEAN predicts an EC number with a confidence score greater than 0.8, the corresponding sequences often have low final pLDDT scores when predicted using AF3.