1
$\begingroup$

I have large number of PDB files of peptides of size 6, 10 and 18. I wanted to cluster the peptide structure based on the secondary structure. I used DSSP in mdtraj module in python to assign DSSP code to backbone residues of each peptide sequence of various sizes of various pdb files. My question is whether it is reliable to use DSSP for peptides?

$\endgroup$
5
  • $\begingroup$ Surely visual inspection will tell you. The PDB provides a viewer. Alpha-helix prediction would be low hanging fruit. $\endgroup$
    – David
    May 28, 2023 at 18:30
  • $\begingroup$ It predicts incorrectly for some of them. That is why I am doubtful $\endgroup$ May 29, 2023 at 17:43
  • 1
    $\begingroup$ Well in a way you know the answer then. But as I remember, DSSP assigns a conformation description to individual residues based on dihedral angle. That is not quite the same as predicting secondary structure for an extended region of a chain. Have you checked the original paper, and can you add an example of incorrect assignment? $\endgroup$
    – David
    May 29, 2023 at 18:13
  • $\begingroup$ They use electrostatic interaction between backbone $C-O --- H-N$ and compare whether the value is less than some cut off value. I guess this creates the problem $\endgroup$ May 29, 2023 at 18:50
  • $\begingroup$ Use an ab initio tool like bioserv.rpbs.univ-paris-diderot.fr/services/PEP-FOLD $\endgroup$
    – pippo1980
    Apr 9 at 4:15

1 Answer 1

-1
$\begingroup$

you certainly can run DSSP as you did but the algorithm is designed for full length proteins but it will work on these short structures too.

$\endgroup$
1
  • $\begingroup$ At the moment your answer is just an assertion, lacking detail or supporting citation. It could be made into something more valuable by explaining the algorithm and what aspects of it cause it to fail with short peptides. $\endgroup$
    – David
    May 29, 2023 at 18:17

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .