How to install Autodock on Ubuntu (Linux) with CUDA GPU support?

Autodock [1] is most widely used for docking. To accelerate the docking process, especially, during virtual screening, Autodock-GPU [2] provides great help. In this article, we are going to install Autodock-GPU on Ubuntu.

Preparing system

Open a terminal by pressing Ctrl+Alt+T. Update and upgrade your system using the following commands:

$ sudo apt-get update

$ sudo apt-get upgrade

Downloading Autodock-GPU

Change to a directory where you want to download the software, let’s say ‘Downloads‘.

$ cd Downloads/

You can either download it from GitHub or past the following command:

$ wget https://github.com/ccsb-scripps/AutoDock-GPU/archive/refs/heads/develop.zip

Installing Autodock-GPU

Unzip the downloaded file and then install using the following command.

$ unzip develop.zip

It will create a new directory, move inside that directory and install.

$ cd AutoDock-GPU-develop/

$ make DEVICE=CUDA NUMWI=128

Here, since we are installing for CUDA, therefore the device name is CUDA. You can write the device name carefully. Other options are GPU, CPU, OCLGPU. NUMWI is the thread block size. Enter these values carefully.

It will take a few minutes to finish the compilation. After that, you will notice a binary, namely, ‘autodock_CUDA_128wi’ in the bin directory.

Now, whenever you want to dock using Autodock-GPU, execute this binary as shown below:

$ ./bin/autodock_CUDA_128wi --ffile protein.mps.fld --lfile ligand.pdbqt --nrun <#runs>

This is the basic command. Read the usage for more options.

NOTE: You might get an error while launching autodock, for instance,

“TclError: no display name and no $DISPLAY environment variable”.

If you encounter this error, then paste the following commands and then run autodock.

$ export DISPLAY=0.0

$ xhost +

This should resolve your problem.

References

1. Morris, G. M., Huey, R., Lindstrom, W., Sanner, M. F., Belew, R. K., Goodsell, D. S., & Olson, A. J. (2009). AutoDock4 and AutoDockTools4: Automated docking with selective receptor flexibility. Journal of computational chemistry, 30(16), 2785-2791.
2. Santos-Martins, D., Solis-Vasquez, L., Tillack, A. F., Sanner, M. F., Koch, A., & Forli, S. (2021). Accelerating AutoDock4 with GPUs and gradient-based local search. Journal of Chemical Theory and Computation, 17(2), 1060-1073.