Annotato pipeline issues

[Antonio-Longo]

Hi again everybody,

I am trying to run the Annotato pipeline. As suggested in the prerequisites, I installed the following:

nextflow v24.04.4.5917
singularity v3.5.3 (go v1.23.0)
conda and mamba (I used the latter to create a dedicated environment)
docker v27.1.1

To test it, I selected a draft MAG genome.fa file (4.5 GB), without any other input data. I so used the following command:

nextflow run main.nf --genome /home/path/to/.fa --species "xxx xxx"

that gave me the attached error and nextflow.log.

I am reading about the suggestion about singularity.pullTimeout (for instance changing it as described here: https://www.nextflow.io/docs/latest/config.html)

Do you have any suggestion,

Thank you and best regards!

Antonio

error.log

[gkanogiannis]

Try with "--singularity.pullTimeout 60m" at the end of your nextflow command

[Antonio-Longo]

still the same error, here the new log:

nextflow.log

thanks

[gkanogiannis]

Ops sorry, try to use single dash, like -singularity.pullTimeout 60m

[gkanogiannis]

Pull the latest version. We removed those mounts that were creating issues.

Edit: it has not been merged yet. Try to manually edit annotation/nextflow/modules/local/funannotate/FunannotatePredict.nf and change line containerOptions = "--bind /env:/env --bind ${params.tmpdir}:/opt/databases --bind /mnt:/mnt" to containerOptions = "--bind ${params.tmpdir}:/opt/databases"

[Antonio-Longo]

almost there!!

is there the possibility of reducing the required gene models to run BUSCO, or I should try with "bigger" MAGs (now it is around 10 MB size)... I found something about --min_training_models

log.txt

[gkanogiannis]

We added a parameter in the pipeline for controlling this value. It is not yet pushed.
@tbrown91 and @diegomics are reviewing the relevant PR.
Until then, you can manually edit annotation/nextflow/modules/local/funannotate/FunannotatePredict.nf and append --min_training_models 150 in the funannotate predict command.

[Antonio-Longo]

So it is finally working! I don't know how to thank you for your time and work. Now i will try it with RNAseq...

Thanks, thanks and thanks again @gkanogiannis

[tbrown91]

Thanks @gkanogiannis for your fix #50

@Antonio-Longo can you pull this version and give it a try?

[Antonio-Longo]

it is working, still some debug to run the pipeline!

[Antonio-Longo]

Hi, i just pulled the pipeline and tried it. Unfortunately it crashed with the following error:

ERROR ~ Failed to publish file: /home/alongo/Desktop/aperti/NEREA/annotato/work/b7/3c1640a510e3a46670eecb5eb3b97f/RM_41.SatOct261202092024; to: /home/alongo/Desktop/aperti/NEREA/annotato/02.Repeatmasker/RepeatModeler/RM_41.SatOct261202092024 [copy] -- See log file for details

-- Check '.nextflow.log' file for details

and from the log it seems that in funannotate to fail:

nextflow.log.txt

I can use -resume as @gkanogiannis suggested, but it would be great to have this working...

Moreover, if i resume the work the in the final .gff3 output does of funannotate is missing the proper annotation (the file reports hypothetical protein for each entry, and if i search one of them on uniprot there is a match). Maybe using a different busco db (but none of the scaffold from the analysis was annotated with default eukaryotes).

Thanks again for your help

[gkanogiannis]

Try maybe --publish_dir_mode symlink

[Antonio-Longo]

ok this is helpful to close the pipeline, but i just noticed that the annotation part of funannotate is missing. Do you think it will be created or am I missing something?

however with interproscan tool i got pretty good results!

thanks again