The input for ADFR includes a receptor and a ligand in PDBQT format, prepared affinity maps, and a set of 3D- fill points that are potential ligand binding regions referred as ‘translational points’. Here we will give a step-by-step procedure to perform flexible receptor docking with ADFR with apo structure of cyclic dependent kinase protein 2 CDK2 (pdb:4EK3) and one of its ligands (pdb:1YKR).
Step 1: Install MGLTools package from Downloads.
Step 2: Download data.zip.
UnZip the file into the directory where you will setup the calculations. The file contained prepared receptor (4EK3_rec.pdbqt) and ligands (reference: 1YKR_lig.pdbqt and input: 1YKR_random.pdbqt) files. The 1YKR_random.pdbqt is the randomized ligand structure provided as input during docking to avoid any bias to the initial ligand pose.
Step 3: Prepare affinity maps and translational points.
This particular ligand overlaps with side chains of the receptor’s apo conformation. Label the receptor residues in the docking box () and zoom in and rotate to observe the overlap of lysine 33 and lysine 89 with the ligand.
Selected side chains are displayed as orange Sticks & Balls.
Receptor side chains to be made flexible can also be selected by typing in the type-in widget.
The docking box fully covers lysine 33 and lysine 89 and the button for flexible side chains validity is now green.
AutoSite identified a single pocket in the box and selected it. The binding pocket fill-points referred as “translational points” are places where ADFR places the ligand root atom (magenta sphere mesh – click the circled anchor icon on the left to show the ligand root atom). The binding pocket fill-points validity button is now green and the “generate maps …” button is now enabled.
Save the maps as 4EK3_rec.zip in the directory where you will setup the calculation.
If for any case you cannot get the zip file, download it here
Step 4: Executing ADFR.
Change to the directory where you want to run the calculation.
Linux and MacOSX:
- open Terminal to execute commands.
- cd $DIR_FOR_RUNNING_CALCULATION
- click on the Start button in the tool bar
- Type cmd<return> to open the Window’s command prompt.
- cd c:\users\<USER_NAME>\$DIR_FOR_RUNNING_CALCULATION
Į $WHERE_YOU_INSTALLED/MGLTools2-latest/bin/adfr 1YKR_random.pdbqt -m 4EK3_rec.zip -r 1YKR_lig.pdbqt --jobName 1YKR_1 --seed -1
Į /Library/MGLTools2/latest/bin/adfr 1YKR_random.pdbqt -m 4EK3_rec.zip -r 1YKR_lig.pdbqt --jobName 1YKR_1 --seed -1
Į "c:\Program Files\MGLTools2-latest\adfr.bat" 1YKR_random.pdbqt -m 4EK3_rec.zip -r 1YKR_lig.pdbqt --jobName 1YKR_1 --seed -1
Unlike AutoDock or Vina, the above-specified commands, will run one GA evolution starting with a random seed and will output one solution. To execute multiple GA evolutions with random seeds, we recommend running the calculation on a cluster where multiple GA evolutions can be run parallel on multiple CPUs.
We will soon provide script to run multiple GA evolutions in your local machine. We have provided a support script to help submitting ADFR jobs in a cluster below.
Running ADFR on clusters - Executing multiple serial jobs.
- Log in to your cluster account.
- Install the installer as described on downloads section.
- ADFR jobs can then be submitted modifying the provided submission script. The script can be used both on PBS as well as SGE queuing systems.
-- Change “$WHERE_YOU_INSTALLED” in the script to point to the path of installation directory.
cmd: > python submitzip.py –q <PBS or SGE> -n <no.of jobs> -l <input_lig.pdbqt> -r <reference_lig.pdbqt> -m <zipFile> -o <output file prefix>
Step 5: Analyzing results from multiple jobs.
The bestScore.py script takes the “<$PATH_TO_SOLNS>” as an input, checks the result directory, analyses the ligand pdbqt outputs from all the jobs and prints out the basename of “<$PATH_TO_SOLNS>”, seed no., lowest energy score and the rmsd with the reference ligand.
Cmd: > python bestscore.py “<$PATH_TO_SOLNS >”.