Hi guys! How are you? I am new to genomeassembler, as well as for Nextflow in general. I would like to know if someone could help me check if I am doing things correctly.

Hi! Sure, what are you trying to do?

So, just to provide context: I have both MGI short-reads and ONT long-reads, but my organism does not have a reference genome. In this case, I would need to left some columns of the samplesheet empty. I need to run nextflow with singularity, and I also have set a params.yml file. So this is how I prepared the samplesheet.csv:

sample,ontreads,hifireads,ref_fasta,ref_gff,shortread_F,shortread_R,paired
HAC_ONT_1,trim_pudu_HAC_ONT_1.fastq.gz,,,,trim_pudu_MGI_1.fastq.gz,trim_pudu_MGI_2.fastq.gz,true
HAC_ONT_2,trim_pudu_HAC_ONT_2.fastq.gz,,,,trim_pudu_MGI_1.fastq.gz,trim_pudu_MGI_2.fastq.gz,true
HAC_ONT_3,trim_pudu_HAC_ONT_3.fastq.gz,,,,trim_pudu_MGI_1.fastq.gz,trim_pudu_MGI_2.fastq.gz,true

Columns that you do not need (i.e. are empty) can also be left out, that might make handling the samplesheet easier.

To first test. then use singularity profile

With

-profile test

you will run a test profile, which is to test the pipeline functionality, and that comes with its own samplesheet, you should do

test,singularity

if you plan on using singularity to see if it works

Hmm oky

Depending on if you want to run on your own computer, or on some HPC, you probably also need a profile for that. For many HPC clusters there are profiles available via nf-core https://github.com/nf-core/configs maybe there is one for your cluster, then you do

-profile test,singularity,profile_name

Ahh I see. Yep, in this case we use a HPC server, I will check that for sure

Computational requirements depend a lot on your data, but, yeah, probably you should use your SLURM cluster, nextflow can interact with SLURM, but usually there are some specifics that may come with the profile (see above)

Great, I will check about the computational setup! Do you see any problems with the params.yml?

But I would not combine

-profile test

with many params, or actualy data, the profile is set to not use many resources and probably will not work with real data

Hmm oky

Ideally you try to get it running with

-profile test

and the config for your cluster

Great! Thanks, Niklas. Have a good day

Thanks, good luck, let me know if anything doesnt work as expected

Hi! How are you? Well, I managed to put everything to run yesterday, but unfortunately now this error is being reported when running FLYE 😕

The issue is probably

--genome-size -1

which is strange

well i did not provide the estimated genome size because this organism has never been sequenced before

It is probably an issue where it is neither provided nor estimated, i will look into it tomorrow

hmm oky, but is this a case in which the pipeline will be stuck or it can keep other processes running to finish analysis?

It is stuck i think

this is my last log

abr.-10 12:18:55.289 [Task monitor] DEBUG n.processor.TaskPollingMonitor - !! executor slurm > tasks to be completed: 2 -- submitted tasks are shown below
~> TaskHandler[jobId: 1144622; id: 3; name: NFCORE_GENOMEASSEMBLER:GENOMEASSEMBLER:PREPARE_SHORTREADS:MERYL_COUNT (HAC_ONT_3); status: RUNNING; exit: -; error: -; workDir: /mnt/beegfs/scratch/mbbdsantana/work/db/cf577eeece003ded526c95c7c37c76 started: 1744243378120; exited: -; ]
~> TaskHandler[jobId: 1144623; id: 2; name: NFCORE_GENOMEASSEMBLER:GENOMEASSEMBLER:PREPARE_SHORTREADS:MERYL_COUNT (HAC_ONT_2); status: RUNNING; exit: -; error: -; workDir: /mnt/beegfs/scratch/mbbdsantana/work/89/501cc65e7c6b3b1ab0459e27dbb8af started: 1744243383074; exited: -; ]

Just provide in bases ;)

Yep, i'll give an estimation based on the closest species

Yes there are other parts that may finish, but flye is the assembler and thus pretty important

yess

To reuse the cached tasks that finished use

-resume

Yep, thanks