Protein Data Bank

The N-O-O-N-I-F-I-C-A-T-I-O-N.

Introduction to Python debugging with PDB.

August 25, 2022.

August 25, 2022.

Like people did a lot of X-ray diffraction.

And if you look at PDB, there's a lot of crystallography data, which is great, which gives you electron densities.

And then when people missed out those electron densities, they were like, they can solve the structure.

So all of those regions are missing.

Mm-hmm.

And I think when we see, when we see structural model predictions and folks say, "Oh, well, you know, I've got this time split on the PDB," for example.

And, you know, I, you know, our model's never seen these molec-these molecules or this protein before, and then hopefully we can, you know, kind of use this understanding to make reasonable folds.

Like that's, that's very much in the spirit of, of being able to make the predictions on this stuff, 'cause no one has seen these mole-- especially our internal molecules, like no one's seen those before.

Mm-hmm.

Um, one thing that was very validating for us is, uh, one of our test sets was a, ano- like, another, uh, orthogonal chemical library that our models had not seen that we had used internally, but it was a copy of a chemical library that had been out in the literature.

And so there was one target in particular, um, it was called, it's a protein, it's called BRD4, where there's an identical molecule from our screen that was also in the PDB, the Protein Data Bank, training set that Bolts had almost certainly seen.

Our models did kinda bad on that molecule.

And what goes into the process is not just data or modeling, it's also about sort of, uh, expertise.

So the example, uh, that, uh, that I have is like with, uh, with AlphaFold, um, we looked at what was possible, uh, with current existing, uh, datasets, because we did not have the, the core expertise of now sort of going to, um, or even the resources, by the way, to, uh, sort of say, let's augment the PDB, uh, by a significant sort of order of magnitude.

It was just not...

I mean, the, the amount of sort of investment that organizations and res-- and scientists across the world had put in to construct that data was invaluable.

Uh, they, their shape might change depending on the context.

And like what AlphaFold did, and in fact, like John, John and I, John Jumper and I basically used to sort of discuss this, like, what are we trying to solve? Uh, we, we are-- we don't know what is the actual true ground state that proteins take and what is the actual distribution of structure that proteins take.

What we are trying to do is basically someone got a structure, deposited it in P- the PDB, and we are trying to replicate and get the same structure.

That's what we did, right? And it just so happens that it's useful.

They're always improving, but there have been a number of issues around the traceability, the quality of the data, mis-annotations, things that have been assigned to the wrong taxonomy.

The genome assemblies themselves, there have been a number of publications that have identified that there's human genomics DNA that have contaminated a number of bacterial genome assemblies And so the quality of that data is really not at the same level of quality that the protein data bank is at.

So if, when I think about digital biology as a field that is really trying to leverage this vast amount of genomics reference data that's out there, the first thing that kind of comes to my mind is that the underlying data itself is not necessarily trustworthy.

You have to spend a tremendous amount of time cleaning it up, curating it, vetting it, figuring out what data sets actually are useful and are not contaminated and have the right quality metrics, et cetera.