Software For Hydrogen Bond Detection In Pdb Data?

Question

[Douglas S. Stones]

In order to study the network structure of proteins, I'm looking for software that can predict the existence of hydrogen bonds in proteins from PDB data. I'm intending on running this software for ~1000 PDB files.

After searching around with Google, there appears to be a range of software available for detecting H-bonds according to varying criteria. I have no real way to distinguish between these.

Ideally, I'm looking for something that is "safe", e.g. well-known and widely used, used by experts, and the definition it implements is fairly mainstream and uncontroversial. I.e. I'd like to minimize the possibility of others questioning the validity of my results based on the choice of software. (Also, ideally, I'd like it to be free.)

Question: What is a "safe" package for detecting hydrogen bonds in proteins from PDB data?

[mikmaksi]

Thanks for asking this. I was actually looking for the same thing just a couple of days ago.

Answer 1

[Neilfws]

First, a quick note: hydrogen bonds are only ever inferred by software, not detected. It's not really a complex procedure; most software simply looks for donor + acceptor pairs separated by an appropriate distance and bond angle.

Two well-known, widely-used tools are:

• DSSP
• STRIDE

In my experience they are pretty similar in terms of performance and operation; STRIDE output is somewhat easier to parse.

Answer 2

[Egon Willighagen]

Douglas, I would recommend Jmol (Open Source). Bob implemented the DSSP algorithm, and Jmol has a such a large user base, that any irregularities have been detected by now. This is the scripting command listed there to calculate hydrogen bonding patterns with DSSP:

calculate hBonds structure

The scripting documentation provides more detail.

Answer 3

[Keith Callenberg]

The HBonds plugin in VMD is another option that is widely used. It has several different parameters that allow you to more finely tune what you want to classify as a hydrogen bond.

You can run scripts in VMD by starting it with the command:

vmd -dispdev text -e yourscript.tcl -eofexit

In your TCL script you could include something like:

measure hbonds 3.5 30 $sel1 $sel2

Where 3.5 is the cut off distance and 30 is the cut off angle, $sel1 is a selection with the hydrogen bond donors, $sel2 is a selection with the hydrogen bond acceptors.

There is also a PyMOL script here that would be good for a large study: http://pldserver1.biochem.queensu.ca/~rlc/work/pymol/list_hbonds.py

Answer 4

[Jan Kosinski]

I think that HBPLUS is a safe bet. Cited many times, also by large scale analyses that needed to detect hydrogen bonds in various PDB datasets.

I think that you can get the latest version within LIGPLOT package