VSPerformanceStatistic workflow

Back to workflows browser


This workflow incorporates the experimental status of a compound, i.e. binder or non-binder, the dockingscore of the compound in each receptor ensemble member, and returns the ensemble best able to discriminateknown binders from known non-binders. Two CSV input files are required for this workflow becauseout of the four algorithms used for the statistical calculations, two require nonbonding eventsto be represented as 0 and the other two require them to be represented as 9999999. For more informationabout this workflow, please refer back to "Progress towards automated Kepler scientific workflowsfor computer-aided drug discovery and molecular simulations" section 3.7.

Required Software

Kepler 2.4 or above
MATLAB R2013b or above
MATLAB Parallelizing license (optional)




There are two ways to run this workflow. One is through the graphic interface. After the workflow is opened,users can double-click the parameters to change their values and press the play button to start. Another wayis to run Kepler via a command line.


d This is the directory for the input CSV files and the output files.

n Number of ensemble receptor structures. For example, if user wants the best 2ensemble receptor structures out of the 8 structures they used for virtualscreening, they will specify 2.

high Name for the CSV file that contains 999999 to represent non-binding events.

zero Name for the CSV file that contains 0 to represent non-binding events.