Biology Reference
In-Depth Information
CXCR4, pdb-id 3OE9). It seems that the main problem here is not the sampling of
possible docking poses but more on the scoring function.
5.1.2
Practical aspects
Most of the molecular modeling methods can be run using free web servers and this is
also true for protein-protein docking methods. Unfortunately, not all the methods are
available via web and some other methods are not fully implemented in web servers
(e.g., Rosetta server (
Schueler-Furman, Wang, & Baker, 2005
) is only carrying out
local docking when using free web server). Due to this, it is usually a good idea to
install the software on a local computer. The best possible environment in the sense
of software availability is a linux-based system, as almost all systems can be used
under linux. The problem is that several noncomputer-oriented scientists may find
linux a little bit challenging environment and also in many organizations linux is typ-
ically not supported. If for some reasons, linux is not an option, one can, in most
cases, use Mac OS X without any major problems (most of the software packages
are also available for Mac) and also Windows-based systems are usually reasonable
well supported, although clearly with smaller number of available software pack-
ages. Based on our own experience, none of these environments is perfect and thus
in our laboratories all operating systems are in use.
Usually most of the software packages are available from the developer web page
as ready compiled files, and these are basically always the best option (e.g.,
GRAMM-X (
Tovchigrechko & Vakser, 2006
) is available as linux and Windows-
ready packages; see
http://vakser.bioinformatics.ku.edu/main/resources_gramm1.
03.php
). Only if truly needed (like in the case of RosettaCommons (
Leaver-Fay
et al., 2011
), see
www.rosettacommons.org/home
), one needs to compile the soft-
ware. In those cases, it is a best practice to first read the Readme.txt file and follow
the instructions as accurately as possible.
Once the software is properly installed, protein docking experiments can be car-
ried out. For all cases, it is of utmost importance to first prepare the files needed for
protein-protein docking. Typically pdb-file format is used as input files for software.
Unfortunately, there are several versions of pdb-file and most programs require that
the input file is processed properly. The best approach is to manually edit the pdb file
and to remove all unnecessary parts, basically leaving only protein residue and atom
coordinates plus possible heteroatom data. One should search the current tutorials for
given software packages to be sure how the file preparation should be carried out.
Another topic to be checked is the effect of possible ligands/cofactors and ion-
ization of charged residues for docking. Again, this is quite software-dependent topic
and it must be taken into account. In general, ligands that are needed for protein-
protein docking should be kept within the input. Also, most of the docking methods
do not require explicit information for ionization of protein residues, although this is
not the case with ligands/cofactors where this data usually must be explicitly defined.
A specific question with GPCR-GPCR docking is the nature of possible com-
plex. It is clear that the interaction will happen within the cell membrane and thus
