Pipeline fail at REPEATMASKER_STAGELIB

[elizabethmcd]

Description of the bug

I am running the most current version of @marchoeppner dev branch on Tower on a real dataset. The workflow passed repeat masking and produced the consensi.fa file but failed very quickly at the subsequent staging process with:

cp: can't stat '/.nextflow/assets/marchoeppner/genomeannotator/assets/repeatmasker/my_genome.fa': No such file or directory

Since I ran this on Tower it's a little difficult to debug things in the .nextflow/assets directory so this isn't obvious to me yet where this my_genome.fa file was supposed to come from.

Command used and terminal output

nextflow run 'https://github.com/marchoeppner/genomeannotator' \
		 -name tick_annotation_test \
		 -params-file 'https://api.tower.nf/ephemeral/zSnWhzERlz_Z4e-E6h2gng.json' \
		 -with-tower \
		 -r dev \
		 -profile docker
Params:
pri_prot_target = 5
custom_config_base = https://raw.githubusercontent.com/nf-core/configs/master
evm_weights = /.nextflow/assets/marchoeppner/genomeannotator/assets/evm/weights.txt
plaintext_email = false
monochrome_logs = false
pasa_nmodels = 1000
spaln_q = 5
max_cpus = 16
spaln_taxon = ixodscap
pri_rnaseq = 4
max_multiqc_email_size = 25.MB
max_time = 240.h
aug_chunk_length = 3000000
nevm = 10
tracedir = ${params.outdir}/pipeline_info
validate_params = true
npart_size = 200000000
max_intron_size = 20000
aug_species = human
pri_wiggle = 2
version = false
spaln_options = -M
publish_dir_mode = copy
aug_training = false
t_prot = P
t_est = E
pasa_config_file = /.nextflow/assets/marchoeppner/genomeannotator/assets/pasa/alignAssembly.config
min_prot_length = 35
aug_options = --alternatives-from-evidence=on --minexonintronprob=0.08 --minmeanexonintronprob=0.4 --maxtracks=3
multiqc_title = test run
pri_trans = 4
pasa = false
custom_config_version = master
max_memory = 128.GB
dummy_gff = /.nextflow/assets/marchoeppner/genomeannotator/assets/empty.gff3
spaln_protein_id_targeted = 90
nproteins = 200
proteins = s3://nf-hifi2genome/tick_annotation/2023-08-25-all-tick-species-proteins.fasta
assembly = s3://nf-hifi2genome/tick_annotation/Arcadia_Amblyomma_americanum_asm001_purged_cleanedup1.fasta
pri_est = 4
pri_prot = 3
min_contig_size = 5000
evm = false
email = elizabeth.mcdaniel@arcadiascience.com
schema_ignore_params = genomes,saveReference,igenomes_base,igenomes_ignore,enable_conda,save-reference
genomes:
pasa_aligner = minimap2
ncrna = false
igenomes_ignore = true
outdir = s3://nf-hifi2genome/tick_annotation/genomeannotator
help = false
show_hidden_params = false
t_rnaseq = E
rm_db = https://www.dfam.org/releases/Dfam_3.5/families/Dfam_curatedonly.h5.gz
busco_lineage = arthropoda_odb10
trinity = false
spaln_protein_id = 60

Relevant files

nf-2ZIBaSqCJUU5ME.log

System information

Run on Tower with Nextflow version 23.04.3, using Docker profile

[marchoeppner]

Thanks for the report. Well, the good news is that the missing file is probably exactly where it is supposed to be ;)

But I am guessing that the combination of Docker/Tower makes assumptions that my way of staging is violating (the woe of having been developing this for such along time...). Basically, it is not a very Nextflow'y way of doing things:

cp ${baseDir}/assets/repeatmasker/my_genome.fa .
is where I think it breaks.

Give me 1-2 days to re-write this into something a bit more up-to-date... (that said, I have no access to Tower or Docker, so we might be poking around in the dark for a little bit).

[elizabethmcd]

Just a small update - I went ahead and ran this on a beefy EC2 instance to see if it would work and everything completed successfully! So I think just a weird way with how Tower is interacting with assets

[rvijay-seqcoast]

Hello, I am running into this same issue using Tower and AWS. Is there a fix or workaround for this problem?

[marchoeppner]

Thanks for your interest in this project. I am sorry to say that I have left research, so the pipeline is - as it stands right now - no longer being maintained. Maybe someone from the nf-core team will eventually decide to either take it over or remove it from the project.