Job Submission Problems

The SBGrid Science Portal is experiencing problems with job submission; please contact if you would like to be notified when this has been resolved

den image

DEN/CNS - Deformable Elastic Network

The DEN portal is a free (after registration) portal for academics wishing to run the CNS/DEN refinement routine. The DEN portal provides access to a grid service provided by the OSG project.

Refinement for low resolution X-ray data

D. J. O'Donovan, I. Stokes-Rees, Y. Nam, S. C. Blacklow, G. F. Schröder, A. T. Brunger and P. Sliz A grid-enabled web service for low-resolution crystal structure refinement Acta Cryst. (2012). D68, 261-267.

Schröder GF, Levitt M, Brunger AT. Super-resolution biomolecular crystallography with low-resolution data. Nature. 2010; 464 (April).

Schröder GF, Brunger AT, Levitt M. Combining efficient conformational sampling with a deformable elastic network model facilitates structure refinement at low resolution. Structure (London, England : 1993). 2007; 15(12) : 1630-41.


At low resolution the number of experimental observables is usually smaller than the number of parameters. Overfitting thus becomes a major issue.

The main of the DEN approach is to refine only those degrees of freedom for which the data provide information. For those degrees of freedom that are not defined by the data we use prior structural information. More information is available from the authors here.


CNS file This grid portal implements a massively parallel parameter optimisation search. You must submit a CNS DEN input file in the 'CNS file' field, and this file must be based on the CNS template `refine_lowres.inp' or `refine_den.inp' which can be found on the CNS website.

Data file The CNS job described in the input file will rely on several data files relevant to your structure that aren't in the CNS library. These additional files (those not prepended in the CNS file with `CNS_TOPPAR:') must be submitted as a gzipped tarball in the `Data file' field of this form. If you are concerned about the privacy of your data, please read our privacy policy below.

The grid will run CNS/DEN jobs for a range of starting positions, DEN weights and values gamma (see paper) and then score the results by rFree, r and the results of a PROCHECK run. The user is then presented with these results and can easily find the 'optimum' parameters for which the CNS / DEN jobs should be run at.

Optimise over temperature? This will vary the starting temperature of the CNS/DEN jobs, allowing the user to also find the optimum starting temperature. Please note that this will cause the grid to run five times more jobs, and so may take five times longer to run.

Include control This will include DEN weight = 0 CNS runs - effectively a CNS refinement without DEN. This is useful for benchmarking DEN and showing how it (may) improve the results of the CNS run.

How many random starting positions The atoms choose (randomly) by CNS for the DEN refinement can drastically effect the efficiency and effectiveness of the refinement. We have independently found ~20 to be a sweet spot, but the user can adjust this as they choose. We have artificially limited the number of possible random starting positions to 32 out of consideration for other grid users.


In order for a job to be accepted, low level consistency checks will be performed on the data attached to the form in order to ensure that all data resembles a correct DEN / CNS job.

Once a job has been accepted a pilot job will be launched on the grid. This is to check that CNS can interpret your job file will return a sensible result. Depending on how busy the grid is and how large your structure is, this can take 12 hours.

When the pilot job has completed without error, then the DEN job proper will be submitted. As this takes a significant amount of grid resources this can take more than a day to complete.


Privacy Policy

Your data and results are secure and can only be accessed by individuals with whom you share the access password. Internally we may review your data and results for the purposes of improving the DEN system. Any future publications on our part will contact owners of the data which we wish to reference for prior approval.