ghm17 / LOGODetect

LOGODetect is a powerful tool to identify small segments that harbor local genetic correlation between two traits/diseases.
GNU General Public License v3.0
19 stars 5 forks source link

No Output #23

Closed NRRdy closed 1 year ago

NRRdy commented 1 year ago

Hello,

I attached a test script I ran using the sample data that came with the installation and following the example outlined in the GitHub (https://github.com/ghm17/LOGODetect#a-concrete-example).

The script appears to have run successfully (see attached log), but in the output folder the program generated a folder called “tmp_files”, but did not generate the desired output (https://github.com/ghm17/LOGODetect#output).

TestScriptLog.txt

LogoDetect_TestScript.txt

Thanks

ghm17 commented 1 year ago

Yes, the script should run successfully based on the log file. Can you show me what's in the tmp_files so that I can check the script stops in what stage?

NRRdy commented 1 year ago

Sure. See attached screen shots. ss1 ss3 ss2

ghm17 commented 1 year ago

Under the '/tmp_files/block' directory, can you find all these files: 'dat1block@.txt', 'dat2block@.txt', 'ldscblock@.txt', 'Qmaxblock@.txt', 'ref_bedblock@.txt', 'ref_bimblock@.txt', 'sd1block@.txt', 'sd2block@.txt' for blocks = 1..204?

NRRdy commented 1 year ago

Yes, those are all there.

Also: Qmax_block1...197 ref_bed_block_1...204 ref_bim_block_1...204 sd1_block1...197 sd2_block1...197

ghm17 commented 1 year ago

Qmax_block should be 1...204 too. Is it possible that your submitted job has been killed due to time limit, and the script has generated files up to 197?

NRRdy commented 1 year ago

I am currently in communication with our IT professional who maintains our server, so I have passed along your question and will get back to you.

This is my first time using LOGODetect. How long does it typically take a job to run?

ghm17 commented 1 year ago

It takes about 15 hours on the example data with 20 cores.

NRRdy commented 1 year ago

Thank you. Yes this issue has been solved by our server admin and was most likely due to a timeout issue like you suggested. This issue is resolved thank you! I do have a couple other questions that aren't really relevant to this thread. Is there another forum to post on or outlet to reach you?

ghm17 commented 1 year ago

No problem! You can contact me at hmguo@stanford.edu.