>So the criticism towards AlphaFold 2 will likely still apply? For example, it’s more accurate for predicting structures similar to existing ones, and fails at novel patterns?
Yes, and there is simply no way to bridge that gap with this technique. We can make it better and better at pattern matching, but it is not going to predict novel folds.
Yes, and there is simply no way to bridge that gap with this technique. We can make it better and better at pattern matching, but it is not going to predict novel folds.