Magnet Programs Overview

The simplest way to use magnet is as a filter to select docked ligand poses which satisfy a set of rules which specify particular geometrical features of ligand-protein structures. By using the keep and reject actions of the SEA rules language a subset of the docked poses can be filtered out of the docking program output and saved. The simplified process flow would look like:

Magnet accepts structure input using the CEX language (which is also used by DockIt). However, it is possible to convert various other standard molecular file formats into CEX using the included babelcx program (a version of the open source babel program with CEX I/O added) so it is possible to use magnet with other docking programs as well.

More commonly, magnet is used to compute properties for particular protein-ligand interactions and output a text file. The output will then be examined and filtered to give a list of ligands and poses which then can used to select the desired ligands and poses. The first step:

This gives a text file which typically has the ligand identifier and pose number in the first column (which will be used in subsequent steps) followed by additional columns of properties (for example, the count of particular hydrogen bonds or other interactions). Then the text file is processed to select a subset of the poses. This can be done by the program magfilter which provides a spreadsheet like functionality or by using standard Unix tools like awk and sort or by using a general purpose spreadsheet program. The resulting text file with ligand names and pose numbers in the first column can then be used by the efcex program to filter the original docking file to select the desired ligands and poses.

An example of this process is provided as a tutorial in the magnet release. More details on the individual programs are available here.