Working with Analysis Results  

When you run a module or pipeline, GenePattern runs the analysis job on the GenePattern server. Analysis results are stored on the GenePattern server for a period of time (by default, one week) and then deleted. If you are unfamiliar with how GenePattern runs modules and pipelines, see Concepts.

• Basic Operations
• Analysis Result Files
• Job Status Page
• Job Results Summary Page
• Sharing Analysis Results

Basic Operations

The following table summarizes ways to work with analysis results:

Analysis Result Files

When you run a module or pipeline, the files generated by the module/pipeline are stored on the GenePattern server. The module author determines the content and format of the generated files; however, by convention, each module generates the following files:

• Analysis result files are typically formatted text files that contain the results of the analysis. Most analysis result files are intended to be used as input to subsequent analyses. Although these files can be viewed in a text viewer, the amount of information in the files may make them difficult to read.
  If an analysis module generates an analysis result file that requires examination and interpretation, there is generally a corresponding visualization module that you can use to display the results. Visualization modules have “Viewer” in the title. For example:
  • Results from ComparativeMarkerSelection are viewed using the ComparativeMarkerSelectionViewer
  • Results from modules in the Prediction category are viewed using the PredictionResultsViewer
  • Results from SOMClustering are viewed using the SOMClusteringViewer
  • Gene expression datasets are often viewed using the HeatMapViewer
  • Large, integrated datasets are often viewed using Integrative Genomics Viewer (IGV) — a high-performance visualization tool that supports a wide variety of data types including sequence alignments, microarrays, and genomic annotations (see the In-Depth Article: Using IGV Through Genepattern)
• Execution log files are text files that describe how the analysis was run. A module might generate one or more of the following log files:
  • gp_task_execution_log.txt: contains the parameter values used to run the analysis, which is useful for reproducing analysis results.
  • stdout.txt: contains "standard output" messages; that is, comments generated as the analysis module runs.
  • stderr.txt: contains "standard error" messages; that is, information about errors (if any) that occurred during the analysis.
  Log files can be viewed using a text viewer.

Job Status Page

When you run a module or pipeline, GenePattern sends the analysis job to the server and displays the Job Status page. This page displays complete information for an analysis job, including its status, input files, parameter values, and (when the job completes) result files. After starting an analysis, you can continue working. You do not have to leave the Job Status page displayed.

GenePattern offers several ways to redisplay a Job Status page:

• Click Modules & Pipelines to display the GenePattern home page, where your most recent jobs are listed in the Recent Jobs pane. Click a job ID number to display the corresponding Job Status page.
• Click Job Results>Results Summary to display all of your analysis jobs. Click a job ID number to display the corresponding Job Status page.
• Enter the URL of the Job Status page in your browser: http://<server>/gp/jobResults/<job number>
  On the GenePattern public server for example: http://genepattern.broadinstitute.org/gp/jobResults/111
  On a local server for example: http://localhost:8080/gp/jobResults/111

The following figure shows the Job Status page for a pipeline job. The Job Status page for a module job is similar.

	Click the menu icon to next to the job name or a file name to display the Job or File menu.
	Icons indicate whether this is an input file or an output file.
	Click the visualizer icon to open a visualizer, if the job includes visualizers.
	Show/hide execution log files.
	Click the share icon to share analysis results with other users. An open share icon indicates that the results are shared.
	Click the parameters icon to display the job's input parameters. To run the job with different parameters, select Reload from the Job menu, enter the desired parameters, and rerun the job.
	Icons indicate whether the job is running, complete, or halted due to an error.
	For a pipeline, each section of the colored line beneath the job name represents a step in the pipeline. As each step completes, its section of the line changes from green to blue.

An Email Reminder check box is visible while the job is running. For long running jobs, select the check box to have GenePattern send you email when the job completes. Continue working in GenePattern or exit from GenePattern. When you receive the email indicating that the job is finished, display the Job Status page to review the analysis results.

Job Results Summary Page

The GenePattern home page lists your most recent jobs. The Job Results Summary page lists all of your analysis jobs.

To display the Job Results Summary page, click Job Results>Results Summary. If a job is still running, refresh the page periodically to update its status.

To sort the job results, click a column header. You can sort jobs by status, job ID, module name, submit date, or completion date. Within jobs, you can sort files by file size or file output date.

	Filter the display: My job results (default) lists all analysis jobs that you have run. All job results lists analysis jobs that you have access to, including jobs that you have run, jobs shared with groups to which you belong, and jobs shared as public. If you are an administrator, you have access to all jobs run on the server so all jobs are displayed. Public job results lists all jobs shared as public. In group: group-name lists all jobs shared with the named group. There is an In Group entry for each group to which you belong. Batch: number lists all jobs run as a result of submitting this batch process.
	Show/hide the execution log files.
	Icons indicate whether the job isrunning, complete, or halted due to an error.
	Click the job ID to display the Job Status page.
	Delete jobs and/or files: (1) select the check boxes of the jobs and/or files to delete and (2) click the delete link in the column header to delete them. Selecting a job selects all of its files. Selecting the check box in the column header selects/clears all check boxes.
	Name of the module that was run and the name of each result file. Click the arrow next to the Module Name header to hide/show all result files. Click the arrow next to a module name to hide/show its result files.
	Size of each file and the total job size (combined file size).
	Time the job was submitted.
	Time the job was completed and time each file was last saved.
	Name of the person who ran the job. Or, more precisely, the GenePattern user name of the account that ran the job.
	Your access to the job. You have read/write access to jobs that you have run. You have either read or read/write access to shared jobs. Write access gives you permission to delete a job or any of its result files.
	Share status: private or shared . By default, jobs that you run are private: only you (and GenePattern administrators) can view or delete them. To share a job, click its job ID. When GenePattern displays the associated Job Status page, click the share icon to share the job.

Sharing Analysis Results

When you run an analysis job, by default, it is private: only you (and GenePattern administrators) can view or delete the job. Sharing job results gives other GenePattern users access to the job, including its input files, parameter values, and result files.

To share job results or modify the share status of a job:

1. Display the Job Status page for the job. The share icon indicates whether the job is private or shared .
2. Click the share icon. GenePattern displays the share options for the job:
   
3. Modify the options by choosing which groups should have access and what access they should have.
   • Which groups have access. Jobs can be shared with all GenePattern users (Public) or groups of GenePattern users. You must be a member of group to share a job with that group; therefore, the options include only groups of which you are a member.
   • What access. Groups can be given three levels of access: None, Read Only, Read and Write. Users with read access to a job can download it, rerun it, and view its input parameters, input files, and result files. Users with write access to a job can delete the job or any of its result files.
4. Click Update Settings to save your changes.

Sharing input files: In GenePattern, you can specify the output file from one analysis as the input file for a subsequent analysis. For example, you might use the output file from PreprocessDataset as the input file for ComparativeMarkerSelection. In this case, if you share the ComparativeMarkerSelection job, the other user can view the result files but cannot view the input file (which is from the PreprocessDataset job) or rerun the job. To share the ComparativeMarkerSelection job and its input file, either (1) share both the ComparativeMarkerSelection and PreprocessDataset jobs or (2) save the output file from PreprocessDataset, rerun ComparativeMarkerSelection using the saved file, and share the resulting ComparativeMarkerSelection job.

Creating groups: To create a group or add members to a group, contact the GenePattern administrator. If you are an administrator, see Creating Groups and Administrators for more information.

<< Running Modules and Pipelines | Up | Working with Modules >>