Issue running GISTIC_2.0 module

[Fernando RB]

Hello everyone,

I've encountered an issue while attempting to run the GISTIC_2.0 module in GenePattern. Specifically, when attempting to run the latest version (6.15.30) on the segmentation file,  the job abruptly terminates with an error. Here is the error message:

"""

INFO: Using cached SIF image Error using ghostscript (line 186) Problem converting PostScript. System returned error: -1.Failed to convert to output format; Ghostscript status: -100.**** Unable to open the initial device, quitting. **** Unable to open the initial device, quitting. Artifex Ghostscript 8.54: **** Could not open temporary file /scratch/genepattern/job_28983437/gs_itxHKB Error in print>LocalPrint (line 307) Error in print (line 233) Error in save_current_figure (line 38) Error in plot_snp_score (line 492) Error in gistic_plots (line 219) Error in run_focal_gistic (line 347) Error in run_gistic20 (line 124) Error in run_gistic2_from_seg (line 249) Error in gp_gistic2_from_seg (line 97) MATLAB:ghostscript:ProblemConvertingPostScriptSystemReturnedError

"""

Then, I attempted version 6.15.28, which appears to be stable. It executed successfully with default input parameters. However, when I tried to adjust some input parameters (amplifications and deletions thresholds, and q-value threshold) it still ran with the default parameters, yielding always the same results.

Any idea of the cause of this issue?

Thank you in advance for your help.

[Andrew Cherniack]

Can you share your seg file?

[Fernando RB]

Of course, this is the segmentation file I used as input.

I used the Human_Hg19.mat file as refgene file.

[Andrew Cherniack]

I don't know if this is the issue, but your column headers with periods might be causing an issue.

Try changing your column headers to this.

Sample Chromosome Start End Num_Probes Segment_Mean

[Andrew Cherniack]

Here is a second issue,

Your seg means are all positive numbers.

How are they centered?  Are they log2 values?

[Andrew Cherniack]

Sorry I was wrong about your seg means.  They look OK

[Fernando RB]

I have tried changing the column headers as advised. However, the problem persists: the execution with version 6.15.30 still aborts, and with version 6.15.28, despite specifying different parameter values, the default parameters are utilized.

[Andrew Cherniack]

Can you send examples of the input parameters that are causing errors?

[Axel Künstner]

Hi Andrew,

some problem here with v6.15.30. The previous version (v6.15.28) runs smoothly using the same input data.

Here the error log:

INFO: Using cached SIF image Error using ghostscript (line 186) Problem converting PostScript. System returned error: -1.Failed to convert to output format; Ghostscript status: -100.**** Unable to open the initial device, quitting. **** Unable to open the initial device, quitting. Artifex Ghostscript 8.54: **** Could not open temporary file /scratch/genepattern/job_29098829/gs_K5CKMW Error in print>LocalPrint (line 307) Error in print (line 233) Error in save_current_figure (line 38) Error in plot_snp_score (line 492) Error in gistic_plots (line 219) Error in run_focal_gistic (line 347) Error in run_gistic20 (line 124) Error in run_gistic2_from_seg (line 249) Error in gp_gistic2_from_seg (line 97) MATLAB:ghostscript:ProblemConvertingPostScriptSystemReturnedError

Best,

 Axel

[Fernando RB]

In version 6.15.30, even the default parameters cause the error.

Despite this, in version 6.15.28, I've managed to generate results using the default parameters. However, when attempting to run the same input data with the q-value threshold set to 0.001 (default=0.25) and another run with amplification and deletion thresholds set to 0.3 (default=0.1), the default parameters are persistently employed, resulting in identical outputs.

[Ted Liefeld]

Fernando

this might have been caused by an issue we tracked down in another help request (problem with the $TMPDIR changing on the compute env to /scratch and that directory not being mounted to the container running the analysis).  Can you retry the 6.15.30 runs to see if they work now?

Thanks for your patience and help,

Ted

[Andrew Cherniack]

Fernando,

I just ran your file with a change of q value to 0.001.  It gave two error messages unrelated to GISTIC itself, but everything seems to have been run to completion.

If you are running HG19 Copy number from TCGA SNP array data, I recommend using the attached marker file.

https://drive.google.com/file/d/10A40YdrjtW2I1uE5gIF6wu1EqvUxZpTv/view?usp=sharing

Regards

Andrew

[Fernando RB]

Version 6.15.30 is now working perfectly.

Thank you very much for your help!