Hi again, New attempt, new failure. Again with the nf-core rnaseq pipeline. This time the problem is with Picard Mark Duplicates. Error exit status (1) seems very uninformative, at least it's not something I can figure out with the help of Google... I have 12 samples. The star_salmon folder contains .bam and .bai files for 3 of them and folders for an additional 2. The picard_metrics folder contains files for 3 samples. So it seems like the pipeline ran through for some of the samples. Does this indicate a problem that I can fix somehow or should I just try again with -resume? Appreciate any help.

Often the files in the work directory gives additional clues about what goes wrong.

You might find the troubleshooting docs helpful: https://nf-co.re/docs/usage/troubleshooting/overview

There are >200 folders in the work directory and I have no idea what I'm looking for. The directory connected to the error in the log file (6b/d2...) is empty. I've come to the conclusion that I'm way out of my depth and do not have the patience to try and figure this out on my own. I'm attempting to get help from NBIS.

You can probably get help at one of the drop-ins (https://nbis.se/get-support/talk-to-us). As for the directory here, it's likely not empty but rather has files that you need to ask specially to see (

-a

to

ls

). There are typically files like

.command.out

and

.command.err

there that might contain clues, if not, there's a

.command.sh

with the actual work that one can typically try running manually from that directory and see what happens.

True, I didn't look for hidden files.

🙂

I know about the drop-in, but I think I should outsource this entire analysis. I've learned enough (thanks to NBIS) to be able to understand what someone else has done, but not enough to be able to do it myself.

Sounds like you figured it out! That's 90% of it 🙂