Adding 3D structures of proteins

[maxim-k]

The problem Gene info, including PDB ids comes from https://mygene.info. But for some genes PDB has multiple ids. I.e. for STAT3 mygene.info returns: 5AX3 - Crystal structure of ERK2 complexed with allosteric and ATP-competitive inhibitors 5U5S - Solution structures of Brd2 second bromodomain in complex with stat3 peptide 6NJS - Stat3 Core in complex with compound SD36 6NUQ - Stat3 Core in complex with compound SI109 6QHD - Lysine acetylated and tyrosine phosphorylated STAT3 in a complex with DNA 6TLC - Unphosphorylated human STAT3 in complex with MS3-6 monobody

Solutions

• Avi suggested using Alpha Fold - https://alphafold.ebi.ac.uk/download but it require either downloading the whole archive, or download PDB/CIF files https://alphafold.ebi.ac.uk/entry/P40763. But I'm not sure whether URLs are regular (i.e. https://alphafold.ebi.ac.uk/files/AF-P40763-F1-model_v1.pdb but what if there is v2?)
• Other Avi's suggestions - use Watir to get these URLs, and this to visualize them.
• Meanwhile, AlphaFold predicted proteins have UniProt ids, and iCn3D takes these ids as an input - https://www.ncbi.nlm.nih.gov/Structure/icn3d/full.html?afid=Q9UEF7 (KL structure predicted by AlphaFold)

[u8sand]

Apparently the AlphaFold website itself

uses this component: https://github.com/PDBeurope/pdbe-molstar

I've implemented it as a react component in the pdbe-molstar branch.

It is trivially instantiated with:

      <PDBeMolstarWidget
        style={{ minHeight: 500 }}
        options={{ moleculeId: '2nnu', hideControls: true }}
      />

Not sure how we get the moleculeId though, perhaps it is, in fact, the ids we get out of mygene.info.

[u8sand]

Here is the docs for the options of the PDBeMolstarWidget https://github.com/PDBeurope/pdbe-molstar/wiki/1.-PDBe-Molstar-as-JS-plugin#plugin-parameters-options