Problem with PDB, DSSP and Protein solvent accessibility

Question

[fabri.digua]

Hello to all, I'm a computer science student and I do not know anything about chemistry and bioinformatics ... I have to build a neural network that, starting from the secondary structure of proteins, preaches "solvent accessibility". My starting dataset is this: http://www.princeton.edu/~jzthree/datasets/ICML2014, but my professor told me to also use dssp ... that I have no idea what it is and what it is for. . Can anyone tell me how to get the "solvent accessibility" from the dataset alone? Do I need to get PDB files first and then convert to DSSP?

Excuse my ignorance on the subject, but I am a novice. Thanks in advance.

Answer 1

[fishgolden]

DSSP is this http://swift.cmbi.ru.nl/gv/dssp/

DSSP can calculate secondary structure and solvent accessibility of (residues in) protein using pdb formatted file as input. Please read the introduction.

http://www.princeton.edu/~jzthree/datasets/ICML2014/cullpdb.tar.gz has over 8000 pdb files in it, thus you can calculate those values using them.

Solvent accessibility can be learned from https://en.wikipedia.org/wiki/Accessible_surface_area

Secondary structure can be learned from https://en.wikipedia.org/wiki/Protein_secondary_structure

But it is better to read some textbook.

[fabri.digua]

Ok thanks! Another question: how can i (if i can) link the entries in the others db (es. cullpdb+profile_6133.npy.gz) with the pdbs files?

[fishgolden]

I think it is written in the paper. (This one http://proceedings.mlr.press/v32/zhou14.html ? I'm not sure cuz I haven't read it.)

There are many papers about Secondary Structure or Solvent Accessibility prediction. https://scholar.google.co.jp/scholar?as_ylo=2005&q=secondary+structure+prediction+-RNA+-server Material & Methods section in those papers will be helpful.