ComputBiophys/ProtRAP

Can't you take into account 3D data?

UnixJunkie opened this issue · 2 comments

UnixJunkie commented

Hello,

If you are given a 3D structure solved experimentally (e.g. in the PDB).
Can't you use that to improve your predictions?
With a structure, we will know which AAs are buried and which
have a chance to be solvent-exposed or exposed to the lipid layer.
Very interesting work!
Thanks a lot for making it open-source!

Regards,
Francois.

Luofan-KK commented

Thanks for your issue. Our predictor is designed to predict RLA, RSA and RBSA from sequences. If you have a 3D structure, you can:

  1. Search for its existence in MemProtMD databse, which is also the training set of ProtRAP. Or you can execute your own molecular dynamics simulations following their protocol.
  2. Use OPM server to calculate the orientation of the protein in the lipid bilayer.

In summary, if you already have a 3D structure, there are several ways to get more accurate results.

UnixJunkie commented

Thanks, I did not know about those.
You know your topic.