Entering edit mode
Kath
•
0
@kath-18711
Last seen 5.1 years ago
Dear support,
I got an bgzf_read error while running diffbind dba.count(). I had posted the message here for the reference, could anyone kindly provide any advises? Thank you in advance!
Hi,
Tagging your question with "software error" is not going to help you. Also you didn't provide your
sessionInfo()
or a reproducible example as required in our Posting Guide. Please make sure to read the Posting Guide if you want to be helped.Thanks, H.
Thanks and here is the sessionInfo()
Good thanks! By having tagged your question with the name of the package you are using you're getting a lot more chance to receive attention and help from the right people. They might still need a reproducible example (a.k.a. "minimal working example" or MWE) in order to be able to help but that's up to them.
How large are the files? Can you try on a non-Windows operating system?
The largest bam is around 2gb. I haven't try mac/linux system yet for diffbind, do you think it's might be due to the platform issue? I can try rstudio at linux and see how it goes then
My guess would be an integer overflow on Windows, the critical size is 'around' 2Gb, specifically
would you be willing to share the problem file?
Thank you! I did try win-based VM linux, and there was no error out while running dba.count(). However, when I plotted the sample correlation in the next step, it's the same plot with those 4 blank samples. It looked that those 4 samples still not readin with dba.count().
I can share those files, where to share?
The update --
I re-run diffbind at linux server instead of the VM, those files processed successfully. It seems that the VM has the same limitation as the windows.
Thank you very much for the comment!