|
The Alphafold team, together with the EBI have released tens of thousands of models of protein structures. Unfortunately, most of these structures did not pass WHAT IF′s LR filter. This website gives some random notes on this problem. |
These pages are not rigorous science. Just a random set of notes and thoughts. These pages are meant to get you started thinking about the problem, and perhaps can be a start for a (very necessary) improvement of Alphafold. After all, the name suggests that this was an α version, after which we expect a β version, and a few years later a production version.
As things stand now, we strongly discourage you from using the EBI-released Alphafold-produced models of proteins.
What′s going on? The next page gives a series of examples of Alphafold models with a very high WHAT IF LR-score.
Perhaps to be removed before we open up this website:
Last year I gave a seminar about appreciation for the application of
artificial intelligence and symmetry operations upon harvesting carrots
by rabbits. This year I talked about the usefulness of AI-derived homology
models for application in molecular science.
These two seminars came out very similar.
:-)