Workflow Builder

The Workflow Builder takes advantage of JMP Genomics' ability to save settings (see Saving and Loading Settings ) to enable you to string together groups of processes into a comprehensive analysis series. A typical, well-designed workflow can take you all the way from the initial data import, through preliminary analysis, quality control, statistical analysis, modeling and annotation of results. The workflow series can be saved for use again and again for different data sets, so long as the basic experimental design remains constant. Existing workflows can also be edited as desired to accommodate changes in experimental design.

The basic schematic for a workflow with N settings is shown below.

Each setting consists of a process, along with its associated parameters (input filenames, specific options, and output filenames). The output files from each setting are used as input for each subsequent setting. The complete output (graphical, tabular, and file) from each setting is written to the workflow journal . Thus, the workflow journal contains output from the entire workflow.

What do I need?

To run the Workflow Builder , you first need to decide:

•	which specific processes you want to run as a group, and

•	in what order to run those processes.

Each process needs its own Input Data Set , which might or might not have been created by a previous process in the workflow. If the workflow requires an import of raw data, the appropriate data import engine as well as each of the processes to be used in the workflow must be configured, and the settings saved in one settings folder. Finally, an output folder must be created, into which all of the resulting data sets, analyses, graphics, and other output are placed.

A sample workflow, in the process of construction, is shown below. Using the selection and ordering buttons, the workflow currently starts with Affymetrix Input Engine , which is followed by Data Distribution , which is then followed by Data Standardize , and ends with ANOVA . The output from each process becomes input for each subsequent process.

The Workflow Builder is not intended for first time users. Instead, it was designed for power users who have already explored the many options offered through JMP Genomics. It is assumed that you are familiar with the processes, and have settled upon one or more standard protocols for your analyses. You also typically have specific saved settings for each of the processes.

Tip : When saving settings for individual processes, you can now opt to save those settings to a newly saved workflow by checking the Save to Workflow box. If the Workflow Builder is open, the new setting is added to the current workflow. If the Workflow Builder is not open, saving a new setting with this box checked launches the Workflow Builder with the newly saved setting in place.

The set of saved settings that comprise a workflow can be edited either in the dialog for that process or in the Workflow Builder itself. If you are not familiar with the individual processes that you want to use, consult the specific documentation for those Processes for more information.

For detailed information about the files and data sets used or created by JMP Life Sciences software, see Files and Data Sets .

Output/Results

As the workflow runs, the individual processes open sequentially, run, and record their output to the specified journal. Once the workflow has run, the journal opens, as shown in the example below.

From the journal, you can open and examine the specific windows resulting from each of the processes of the workflow. As you evaluate the output of this workflow, you can opt to go back, modify the settings in the individual processes, and rerun the workflow until you are satisfied with the analysis.

Tip : As you become more experienced with workflows, you can apply them directly to new experiments. In this case, it is suggested that you use generic names for all input and output data sets, and use folder names to identify the experiments. That way, the results will not contain data set names from previous experiments.