3 ms·
Shouldn't the next step just be predicting a distribution of possible structures, instead of just the most likely? That doesn't seem crazy given where we're at
by sdenton4 6y ago
Shouldn't the next step just be predicting a distribution of possible structures, instead of just the most likely? That doesn't seem crazy given where we're at on the machine learning side; predicting distributions is almost the default at this point, in many areas.
- Scandiravian 6y agoThat is actually one of the "classic" ways of doing it. There's a bunch of different Bayesian models for protein structure - as an example, I implemented a model for structural changes in evolution for my master's thesis The big challenge for these kind of models is the curse of dimensionality. Since every atom in the structure can potentially interact with every other atom, it's tricky to make a joint distribution for the entire sequence and it's rare to have a model that's both accurate and parallelizable, so the field hasn't benefitted much from the advances in for instance GPU computing