Benjamin Buchfink
Benjamin Buchfink
Makes sense, I'll look into implementing it.
Your block size is 16 GB and the query file is ~11 GB, so it will be one query block.
That looks like a bug and needs to be fixed, try running without -M for now.
This indicates an error in your `--taxonmap` file. Check that is has the correct header line.
The `--mp-init` call needs to be done once on a head node, not within your submit script. You should also include the sensitivity setting there.
I think you need to use `srun` for the diamond call to make it run multiple instances.
This should fix the issue: d700d155f9aa0af60158efa60329f4ad09c2f373
It should be fixed in the latest release.
There is no other way than running the alignment twice. The DAA format is deprecated and will not be developed further. Of course it is possible to take the diamond...
To be more precise: I meant using `diamond view` to convert the DAA to tsv, then joining against the taxonomy mapping.