Forum:Is there a need for a new Protein/Molecule viewer library?

Question

[Elbbard]

Hello,

I am completely new here so forgive me if I wriite anything wrong and please let me know my post doesn't belong on this forum.

I am a software engineer and I found interest recently in protein folding. As I wanted to start on a new project with Rust and WebGPU, I decided to try to make a 3d viewer for proteins/molecules. I have very quickly put together a prototype that loads a PDB file and shows the atoms as spheres.

I am now wondering if it is worth pursuing this project and I wanted to ask some questions:

• Is there any need for a new cross-platform performant 3d viewer for proteins?
• Is there any missing feature in existing 3d viewers?
• Is there something else that could be more useful to develop in your opinion related to biology/chemistry?

Thank you so much for reading!

Answer 1

[Mensur Dlakic]

I think you should familiarize yourself with "the competition" as that might help you answer your own questions.

• https://pymol.org/
• https://www.cgl.ucsf.edu/chimerax/
• https://www.ncbi.nlm.nih.gov/Structure/CN3D/cn3d.shtml
• https://www.ks.uiuc.edu/Research/vmd/

Can you do it better than they do? If so, can you do it with a reasonable time and effort? If the answer is yes to both questions, what do you expect out of this and how would you convince the potential customers to use your program? Keep in mind that most are already entrenched in their choice of 3D viewers.

I am tempted to tell you not to bother with this subject, but here is what's stopping me. PyMol's author, the late Warren DeLano, was also once an outsider in the field who was told not to bother writing a 3D visualization program, especially not in Python. Yet he made one of the most widely used and visually stunning programs. The probability of anyone replicating that feat in the same 30 years is very small, but who knows.

https://en.wikipedia.org/wiki/PyMOL

[Elbbard]

Thank you very much for the detailed answer and for sharing those links. It means a lot to me that you took the time to answer. I agree it makes a lot of sense to first study what’s already out there and see what could realistically be improved. I’ve started looking deeper into the existing tools to understand their strengths and limitations. I think you are right that it might be difficult to do much better than existing solutions. Maybe I could test out those libraries and see if they are easy to integrate into existing software. I could try to make a very simple/powerful API for other developers but I doubt that it would justify a whole new development. I will take some time now to really think about where I could have an impact that is bigger than "just another protein viewer".

Answer 2

[Joe]

Unfortunately I think the answer is no, unless as Mensur said, you think you can offer something different/better (though even then I personally would emplore you to build a plugin/extension or something instead as I _despise_ fractured software ecosystems - there would definitely be a niche for well made, easy to use, powerful extensions to popular tools). For me personally there is already too much choice in protein modelling apps (I am a Chimera fanboy).

The main one you'd have to better would be the PDB native viewer (Mol* and/or JSMol) which is already a web app (and a pretty good one I'd say):

https://www.rcsb.org/3d-view/9DDG/1

If you were building something more than a viewer, then some of the others Mensur mentioned become the competition.

[Elbbard]

Thank you for your perspective! I see your point about not adding to the fractured ecosystem. I have tested the web viewer that you linked and it is already very rich and easy to use. It does seem difficult to make it much better. One limitation of this viewer could be very large proteins. It is probably using WebGL which is an old technology for 3d rendering. Many browsers now support the WebGPU specification which provides much better performances. I'm not sure it justifies making an entirely new viewer just for that reason though. I will try to take some time now to look into what I can make to have a bigger impact than "just another protein viewer".