refresh-bio/KMC

kmc_tools dump not creating file

kzuckerman opened this issue · 3 comments

I am trying to run KMC 3.2.1 (installed from conda biotools) to count kmers of length 12 in the human genome and dump those counts to a text file. The kmc runs fine, but when I run kmc_tools dump nothing happens. I'm using a Macpro 2018 with an intel core, running Ventura. I've tried uninstalling and reinstalling, no luck. It might just be an issue with my computer not the software.

Exact commands are:
kmc -k12 -t6 -fm -cs1000000000 -b human_genome.fna human_12mers .
kmc_tools -t1 transform human_12mers dump human_12mers.txt

This seems to only be happening with 12-mers which is even stranger. On some runs I get
Error: Cannot open file human_12mers.kmc_pre
Sometimes
in1: 19% zsh: segmentation fault kmc_tools -t1 transform dump
but most times don't throw an error and just don't create a file

Hi, thanks!.
This is not good.
Could you try the newest release, i.e. 3.2.2, you may download it from here: https://github.com/refresh-bio/KMC/releases/download/v3.2.2/KMC3.2.2.mac.x64.tar.gz
then unpack with tar -xvf KMC3.2.2.linux.x64.tar.gz and then run as follows:

bin/kmc -k12 <rest of params>

and kmc_tools in a similar way.

Unfortunately, for unknown reasons, I currently do not have access to our Intel mac, so I cannot try it on my own. On M1 it works (I used this ref. genome: https://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/annotation/GRCh38_latest/refseq_identifiers/GRCh38_latest_genomic.fna.gz).

What other values of k have you tried? Have you used anything lower? The thing is small k are treated in a special way which may be related to the issue.

Hi,
I was able to connect to our Intel MAC and I confirm I am able to reproduce this with v.3.2.1 - kmc_tools randomly fails. I am not sure what is the reason :(, but version 3.2.2 worked each time I tried, so I would recommend switching to 3.2.2.

Thanks again for reporting this!

That worked. Thank you!