Running Covalent Docking from the Command Line

Command-line covalent docking jobs can be set up with command options, or with an input file, or both. The syntax of the command is:

covalent_docking [input-file] [options] receptor-file ligand-file attachment-residue

where:

  • input-file is an optional input file with extension .inp. This file can be written out from the Covalent Docking Panel, by choosing Write from the Settings button menu.

  • receptor-file is a Maestro file that contains the receptor structure

  • ligand-file is a Maestro file that contains the ligand structures

  • attachment-residue identifies the residue on the protein to which the ligands are attached, in the format chain:residue-number, where chain is the chain name, and residue-number is the number of the residue in that chain, e.g A:123.

For information on the command options, see covalent_docking Command Help or run the command covalent_docking -h. Standard Job Control options that specify the host and other resources are also accepted. These options are described in Running Jobs From the Command Line. Some features are available from the command line that are not available from the GUI, such as support for Glide constraints and multiple reaction chemistries.

The input file uses exactly the same syntax as the command options, except that the initial dash is not present when you convert the command option into an input keyword. Keywords are case-insensitive, and are separated from their value by spaces. Any options given on the command line supersede the corresponding options in the input file.

As the jobs can take several hours per ligand in pose prediction mode, you should consider distributing them over multiple processors. You can do this with the -NJOBS option. To protect against failure of the subjobs (e.g. due to network issues), you can set the -RETRIES option to specify the number of times the master job reruns a failed subjob before it is considered to have failed. If the job finishes with failed subjobs, or if the job itself fails for any reason, you can restart the job with the -RESTART option, and it will resume at the first subjob that did not run successfully, keeping the results of the successful subjobs.