Comment by adamsmith143
2 years ago
> it can't predict folds that haven't been seen
This seems strange to me. The entire point of these types of models is to predict things on unseen data. Are you saying Deepmind is completely lying about their model?
Deepmind solved CASP, isn't the entire point of that competition to predict unseen structures?
If AlphaFold doesn't predict anything then what are you using it to do?
AlphaFold figures out that my input sequence (which has no structural data) is similar to this other protein that has structural data. Or maybe different parts of different proteins. It does this extremely well.
This is a gross misrepresentation of the method.
Perhaps you'd care to explain how? AlphaFold does not work on new folds. It ultimately relies on mapping sequence to structure. It does it better than anyone else, and in ways a human probably couldn't, but if you give it a brand new fold with no relation to other folds, it cannot predict it. I routinely areas of extremely low confidence many of my AlphaFold models. I work in organisms that have virtually 0 sequence identity. This is a problem I deal with every day. I wish AlphaFold worked in the way you are suggesting, but it just flat out does not.
11 replies →