gbouras13 / hybracter

Automated long-read first bacterial genome assembly tool implemented in Snakemake using Snaketool.
MIT License
106 stars 8 forks source link

[BUG] - kmc error #98

Open suzumar opened 1 month ago

suzumar commented 1 month ago

Hello, I am trying to use hybracter to assemble a genome (the hybrid mode worked fine) but the long single is giving me this pesky error, for which I cannot find a way around. It stops in the kmc step, and based on the I suspect the snakefile is missing to declare a directory creation (see command lauched on prompt at the end, but before I spend a lot of time to figure out how, perhaps you will see something I am doing wrong or missing. I have a screen capture below Thanks!

hybracter long-single -l raw_reads.fastq -s BBCC288 -c 3500000 -t 8 -d /work/databases/hybra --medakaModel r1041_e82_260bps_sup_g632 --flyeModel --nano-raw

Screen Shot 2024-10-07 at 4 27 45 PM

gbouras13 commented 1 month ago

Hi @suzumar ,

Strange error - snakemake will handle directory creation, but it is deleting them on shutdown, so unsurprising you run into that no such file thing in the manual kmc command.

I would assume kmc is erroring in some way, but your command won't find it as you need to activate the specific kmc environment.

To replicate that, you will need to activate the conda env with kmc (in the screenshot)

conda activate ../../../......./02a etc

Then try the kmc command (without the output file) and see what it says.

George

suzumar commented 1 month ago

Hi George, Thanks a lot I did what you suggest, and I got the same error with "missing log file" I created the directories kmc/BBCC288 and KMC now ran, but I got a error different error message, that perhaps you might help interpret. See screen capture : PS: I did a rwx to all, so it is not a permissions issue as far as I can tell

command kmc -k21 -ci10 -t8 -v -fq hybracter_out/processing/qc/BBCC288_sanitised.fastq.gz hybracter_out/processing/kmc/BBCC288/BBCC288 hybracter_out/processing/kmc/BBCC288 > hybracter_out/processing/kmc/BBCC288/BBCC288_kmcLOG.txt 2>&1

Screen Shot 2024-10-08 at 9 21 21 AM

gbouras13 commented 1 month ago

Oh that is really weird, therefore I guess it has to be a KMC issue.

Looks like this is the issue

https://github.com/refresh-bio/KMC/issues/140

So try ulimit -n 2048

George

suzumar commented 1 month ago

Unfortunately that dit not do the trick. Perhaps I should ask for help on the kmc developers github

gbouras13 commented 1 month ago

Perhaps that would be a good idea. Alternatively, if you use Linux, I'd recommend the hybracter docker/singularity container - should ideally solve this issue.

George