Error of missing metaproteomics input despite specifying only metagenomics data

[shaman-narayanasamy]

Dear developers/maintainers/authors,

Thank you for developing and maintaining MOSCA. We are interested in applying your implementation on our multi-omics data set, but have ran into some issues. Please find the details below. Do not hesitate to ask for any further information.

Looking forward to hear from you.

Best wishes,
Shaman

Description

We are attempting to run MOSCA (for now) using only metagenomics as input. We are obtaining the following error:

$ mosca -c config.yaml
MOSCA command: snakemake -s /ibex/user/naras0c/conda-environments/mosca_env/share/MOSCA/Snakefile --printshellcmds --cores 24 --configfile /ibex/project/e3015/fatima/workdir/mosca_output_shaman/config.json --use-conda
Building DAG of jobs...
MissingInputException in rule normalization in file /ibex/user/naras0c/conda-environments/mosca_env/share/MOSCA/rules/normalization.smk, line 1:
Missing input files for rule normalization:
    output: /ibex/project/e3015/fatima/workdir/mosca_output_shaman/Quantification/mg_normalized.tsv, /ibex/project/e3015/fatima/workdir/mosca_output_shaman/Metaproteomics/mp_normalized.tsv
    affected files:
        /ibex/project/e3015/fatima/workdir/mosca_output_shaman/Metaproteomics/mp_entry_quant.tsv

To give you more context, we created the config file using the MOSCA configuration generator, and only selected metagenomics to be run, i.e. metatranscriptomics and metaproteomics de-selected. However, the error clearly shows that Snakemake is somehow expecting metaproteomics as input. I attempted to comment out content related to proteomics in the config file, but it still produced the same error. Could you please advice on how we could rectify this issue?

To Reproduce

config YAML file: config.yaml.zip. Produced using MOSCA configuration generator

Command: $ mosca -c config.yaml

• OS: Linux HPC environment, with slurm scheduling capability
• Version of MOSCA:

$ mosca -v
MOSCA 2.3.0

[iquasere]

Hey there!

Sorry for taking so long to respond. You raised this issue on the last weeks of my PhD.

I really have to update MOSGUITO. Your config file brings some outdated parameters, but these are not causing the problem here. MOSCA should handle MG data alone, but since it was always tested with either MG and MT or MG and MP together, it was always expecting some type of expression data. I've fixed that now, and will launch version 2.3.1 this week with those fixes. I'll post here the new config file, with your parameters.

Also, congrats for IMP! It was actually the first meta-omics pipeline I ever came across, and it inspired my supervisor to start development of such a pipeline, but for expression analysis.

[shaman-narayanasamy]

Hi @iquasere !

No worries. I have been in your shoes at some point in my life :) I am sure you aced the PhD. Congratulations and all the best in your future work!

We will give the new version a try, but not likely anytime soon. Unfortunately, due to deadlines, we ended up implementing and executing our in-house MG pipeline. Perhaps, we could use it when we have new data.

Thanks for the complement on IMP, it was a long time ago. It is now being maintained exclusive for in-house use by my former supervisor. Making software to be used by others, in other infrastructure is far more difficult, and I commend your consistency and effort!