DSXONLINE v0.88 - Scoring Functions for Protein-Ligand Complexes 0 user(s) online
frameMarburg, Sunday, June 07th, 2020
frame Home
frame Introduction
frame Help
frame Contact
frame Login
frame Register
frame Get fconv
frame Get DSX standalone

 Server is IDLE

frame frame

Documentation and Help

  1. Submit a Job to DSXONLINE
  2. View and Download Job Results
  3. Your Job List
  4. Troubleshooting

  1. Submit a Job to DSXONLINE
  2. To successfully submit a job to DSXONLINE, you basically need a protein file and a ligand file. However, on the job submission form, you may choose among several options and additional files:

    I. Input structure files
    • protein: Specify the path to the protein file on your local file system. This may be either a PDB or a MOL2 file. In the PDB-case only ATOM entries will be considered, while in the latter case everything within the MOL2 file will be treated as part of the protein.
    • ligand(s): You may either submit a single ligand or multiple ligands (and/or multiple geometries of one ligand). Valid input formats are MOL2 and DLG (AUTODOCK solutions).
    • reference: If you provide a reference ligand, in addition to the scores also the rmsd between the ligand(s) and the reference is calculated. This is only reasonable, if the submitted ligand geometries display alternative conformers of the original reference ligand. If not, the rmsd of the maximum common substructures will be calculated.
    • cofactor: You may provide a cofactor in a MOL2 file. Here it is important to choose the correct interaction mode. If treated as part of the protein, all MOLECULE entries within the corresponding MOL2 are simply used as if they were protein atoms. If treated seperately, MOLECULE entry X will be used with LIGAND X and interactions protein-ligand, protein-cofactor, cofactor-ligand will be scored. Consequently, in interaction modes 0,5,6,7 the number of MOLECULE entries in the cofactor file must equal the number of ligands in your ligands file!
    • waters: You may provide a MOL2 with water molecules to be considered. For the interaction mode the same holds true as for cofactors.
    • metals: You may provide a MOL2 with metal atoms to be considered. For the interaction mode the same holds true as for cofactors.

    II. Pair potential type
    • CSD: chooses the pair potentials derived from small molecule crystal data as stored in the CSD.
    • PDB: chooses the pair potentials derived from protein-ligand complexes as stored in the PDB.

    III. Interaction mode
    • This option is of no interest as long as you do not supply an additional file with cofactors, waters or metals. First, please see I. Input - cofactor!
    • Example: If you have a cofactor which was kept rigid (as part of the protein) in docking its a good idea to do the same in rescoring, thus using the default interaction mode where your cofactor is treated as part of the protein.
      But if your cofactor was kept flexible within docking, you must also consider cofactor-protein interactions in rescoring and of course you should have a corresponding cofactor pose for each ligand pose.
    • Consequently, it is advisable to use interaction mode 0,2,4 or 6 if you switch on the GOLD-WATER flag!

    IV. Further options
    • SCORE GOLD-WATER: Switch this on if your docking solutions are from a GOLD run considering water molecules. GOLD stores the informations about used water molecules within the MOL2 file of the docking solutions. DrugScoreX will read this information and consider ON-marked water for scoring of the corresponding ligand.
    • SCORE TORSIONS: This option has only an effect if using CSD potentials. The torsion potentials are also statistical potentials derived from CSD structures.
    • SCORE SOLVENT ACCESSIBLE SURFACE: If switched on, solvent accessible surface ratio potentials are used as an additional term.
    • CONSIDER COVALENT LIGANDS: If you have covalently bound ligands you have to switch on this option.
    • VISUALIZE: DrugScoreONLINE visualizes the per-atom score contributions from the distance-dependent pair potentials. Also single atom-atom interactions will be shown if they are below or beyond a distinct threshold. Bad torsions will be shown if the corresponding option was switched on. From your results you have to download the .py visualization file together with the protein-, ligand- and other files you supplied. Put those files in the same directory and start the .py file with Pymol.
    • NOTIFY: If you check this option, you will be notified upon job completion.

    After submission of the structural data, your job request is placed in a queue. The queue is processed from top to bottom, i.e. the longest waiting job is processed next. Once your job is started, the results will be available within seconds (up to minutes when using using SR potentials).

  3. View and Download Job Results
  4. Once a job is finished, a 'results' link appears in your joblist. Following this link, you will find a summary of your selected job options, a table summarizing your scoring results, and a download section.

    You may sort your results by the ligand number, the rmsd, or the score rank by clicking on the table headers.

  5. Your Job List
  6. In the job list you can manage your submitted jobs. You may encounter different status types.

    Status types:
    • ENQUEUED: The server has accepted your job request and will start processing your job when the queue position is reached.
    • STARTED: The server has started your job. Details on the tasks performed can be seen following the 'job details' link.
    • FINISHED: DSXONLINE successfully finished your job. Your results are available following the 'results' link.
    • SUSPENDED: The server administrator stopped queue processing. Your job will be re-queued as soon as the server is processing jobs again.
    • KILLED: An error was encountered while processing your data. View your job details to get an idea at which stage the job was killed. See also the following section Troubleshooting.

    Your jobs will be hold for at least 20 days.

  7. Troubleshooting
    • Some or all scores are 0
      • This is observed, if the input ligand's distance to the protein is greater than 6Å. In most cases, the coordinates of either the ligand or the protein have been centered somehow without regarding the counterpart.
    • ERROR: DSX failed.
      • If you are using an interaction mode other than 1, make sure to have the same number of MOLECULEs in your optional files as in your ligand file.
      • For parsing and atom type assignment, DSX uses the same libraries as fconv. You may check your input files with fconv in order to see if something goes wrong here.
    • I read your troubleshooting section, but I still cannot get it work
      If you encounter errors that you cannot resolve yourself, you may choose to contact us by means of the 'Contact' form. If you want us to investigate your problem, you have to provide your user name and the job id. This enables us to view your user data and to either give details on the reasons that caused the problems or to re-enqueue your job in case that we could resolve the problem.