Supersecondary Structure Prediction

Question

[Jamand]

Dear All,

Do you know any tool to predict supersecondary structure in proteins?

thank in advance.

[Khader Shameer]

What is your data type ? Sequence or Structure ? Are you looking at any specific type of supersecondary structure ?

Answer 1

[Jarretinha]

You can use Threader and its derivates. It can statistically recognize the fold of your protein/domain. It's pretty slow, but quite reliable.

[Khader Shameer]

fold != supersecondary structure. To the best of my knowledge THREADER is a fold recognition program, not a supersecondary structure prediction program.

[Jarretinha]

That's true. Nevertheless, you can use it for SSS prediction if you don't have a target in mind (e. g. transmembrane helices). You just need to divide your sequence in appropriate pieces. It's quite brute force, it's slow but works. Test it yourself using some known pi-helices. Besides this approach, I don't know a generic SSS server/program.

Answer 2

[Khader Shameer]

Not sure if there is a consensus prediction program which can predict various supersecondary structures (for example Helix-Turn-Helix, Beta-hairpins etc.) from sequence or structure information. Structure prediction is an area of active area research and various algorithms are available. I would recommend you to make a list of your supersecondary structures of interest and then search for servers/program which could predict such features from sequence information.

For example:

Helix-Turn-Helix prediction: http://npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_hth.html

Beta-hair pin prediction: http://www.imtech.res.in/raghava/bhairpred/supli.html

Also refer to Chou K. C's publications and various secondary structure prediction servers from G. P. S. Raghava's group for a good background reading on available supersecondary structure prediction algorithms.

[Jarretinha]

Khader, your links proved much more useful than I thought initially! I've caught myself using Raghava's servers regularly by now (and recommending).

Answer 3

[Suk211]

If you have the PDB file for your protein of interest, you can use promotif to calculate the super secondary structures.

Answer 4

[Anuraj Nayarisseri]

Try out the following tools

• http://myhits.isb-sib.ch/cgi-bin/motif_scan
• http://www.ebi.ac.uk/Tools/pfa/iprscan/

These are the best tools

[Khader Shameer]

Anuraj: MotifScan and iprscan is for functional motif prediction. jamAnd is looking for supersecondary structure information, not functional motifs/patterns.