hisat2-build half done

Question

[crystalpang]

I was running hisat2-build on wheat genome. It seems to me the job stopped somehow. I got 7 files, it's clear that the job is half done. How do I know what's wrong in the process? Am I supposed to run the whole process from top again?

[ATpoint]

You should get used to save the logs that the tool produces, e.g. with 2> to collect stderr to a file. This will tell you if something went wrong. In this case 7 and 8 seem suspiciously small. I would re-do it and collect the logs.

[crystalpang]

Thank you. I am redoing it now. It would be perfect if the program can pick up from where it went wrong, the wheat genome is huge and the index building really takes time.

Answer 1

[michael.ante]

Wheat has a huge genome. With STAR I needed more than 145 GB memory and some tricks as described in several post to reduce the Index size.

I don't know what memory footprint Hisat2 is producing, but I guess you have similar memory requirements.

Cheers,

Michael