Set up and run a prediction

*Note: Before initiating a prediction, you must install Docker and the Nova applications, and prepare the input file(s).

Predictions are initiated by typing a script consisting of commands and “arguments,” which are parameters that affect those commands. Commands and arguments are described in detail in separate topics: Required commands, Required arguments, Optional arguments, and Supplemental arguments.

The following step-by-step instructions discuss a generic script:

1. Launch the terminal.

2. Type the relevant application command and its parameters.
   
   --lshost=ipaddress (or =hostname)
   
   In this step, lshost must be in lower-case letters and preceded by two dashes (-).

3. (optional) Set up a persistent environment variable to define the license server address. Configuration varies by Linux distribution and command shell. Please consult your distribution’s documentation to determine the preferred way to set environment variables.
   
   For example, if you are using a Bourne-style shell like <code>bash</code>, you can set the environment variable for your user account by adding the line:
   
   export DNASTAR_LSHOST=<license server IP address>
   
   to one of the following files:
   
   $HOME/.profile
   
   $HOME/.bash_profile
   
   $HOME/.bashrc
   
   Note that DNASTAR_LSHOST must be in capital letters and should not be preceded by any dashes. In addition, no spaces should be entered before or after the equal sign.
   
   To test if the environment variable has been set, open a new shell and run the <code>env</code> command to confirm DNASTAR_LSHOST is listed.

4. (optional) Append any additional run parameters:
   
   novafold --datadir value1 --seqname value2 --seqfile value3
   
   novafold-antibody --datadir value1 - – seqname value2 --seqfile value3
   
   novadock --output-dir value1 --name value2 --lfile value3 --rfile value4
   
   …where:
   

   novafold novafold-antibody novadock	Command to run the associated application. A single Linux machine should only run one prediction and one application at a time since all resources will be used.
   value1	Path to the workspace or output directory.
   value2	Any desired name consisting of alphanumeric characters, hyphens and/or underscores. No spaces are allowed.
   value3	Path to an input file, such as a sequence file or a ligand PDB file.
   value4	Path to an input file, such as a receptor PDB file.

5. (optional) If you did not set up the environment variable in Step 3, above, append the IP address or hostname specification for your DNASTAR License Server:
   
   --LSHOST ipaddress (or hostname)

6. (optional) Append any additional run parameters. An example NovaFold script appears below these steps. For a list of available parameters for all three Nova applications, see Required commands, Required arguments, Optional arguments, and Supplemental arguments.

7. Press Enter to begin the prediction. For NovaFold and NovaDock, the prediction may require several hours up to several days to finish, depending on the input. NovaFold Antibody predictions usually require only about 5-10 minutes. If you wish to stop a prediction that is in progress, see Stop a prediction in progress.

After a prediction has finished successfully, result files will be added to the workspace folder (datadir) or the output directory referenced in the script.

Example script:

novafold --datadir ~/Desktop/BioluminescentSeqs --seqname firefly_luciferase --hours 100 --GO true --EC true --LBS true

The NovaFold script above denotes that the protein sequence ~/Desktop/BioluminescentSeqs/seq.fasta should be used as the query, and each simulation should run for a maximum of 100 hours. Besides the structure prediction, the following optional predictions should be made: protein function GO terms, enzyme active sites, and ligand binding sites. Results are to be saved as ~/Desktop/BioluminescentSeqs/results.novafold.