moreinfo

Hi James — sorry, updating the database to match the nf-core version did not resolve the issue. Here are the full details.

1) Job script

#!/bin/bash -l
#$ -S /bin/bash
#$ -N funcscan_MAG30
#$ -cwd
#$ -pe smp 1
#$ -l h_rt=48:00:00
#$ -l mem=8G
#$ -o logs/funcscan_MAG30.$JOB_ID.out
#$ -e logs/funcscan_MAG30.$JOB_ID.err

set -euo pipefail

module purge
module load java/temurin-17/17.0.2_8

export NXF_OPTS='-Xms1g -Xmx6g'

export NXF_HOME="$HOME/Scratch/.nextflow"
export NXF_TEMP="$HOME/Scratch/.nextflow_tmp"
mkdir -p "$NXF_HOME" "$NXF_TEMP"

export APPTAINER_CACHEDIR="$HOME/Scratch/.apptainer"
export NXF_SINGULARITY_CACHEDIR="$HOME/Scratch/.apptainer/pull"
mkdir -p "$APPTAINER_CACHEDIR" "$NXF_SINGULARITY_CACHEDIR"

nextflow run nf-core/funcscan \
  -r 3.0.0 \
  -profile ucl_myriad \
  -resume \
  -work-dir "$HOME/Scratch/funcscan_MAG30_work" \
  --input samplesheet.csv \
  --outdir results_funcscan_3.0.0 \
  --run_amp_screening \
  --run_arg_screening \
  --run_bgc_screening \
  --run_protein_annotation \
  --protein_annotation_interproscan_db /myriadfs/home/zcbepgl/Scratch/databases/interproscan_5.59/interproscan-5.59-91.0 \
  --protein_annotation_interproscan_applications Pfam,Hamap,PANTHER,TIGRFAM,sfld,ProSiteProfiles,ProSitePatterns \
  --bgc_mincontiglength 3000 \
  --bgc_antismash_contigminlength 3000

2) Samplesheet

sample,fasta
MAG30,/home/zcbepgl/Scratch/MAG30_ipr/MAG30.fa

3) InterProScan process status

The INTERPROSCAN process completes successfully (exit 0):

Process NFCORE_FUNCSCAN:FUNCSCAN:PROTEIN_ANNOTATION:INTERPROSCAN (MAG30); status: COMPLETED; exit: 0

However, the only error i see in the nextflow log is:

Skipping output binding because one or more optional files are missing.

4) Output behaviour Pyrodigal predicts ~6000 proteins, but the InterProScan TSV output appears heavily dominated by PF00004 (AAA ATPase family) and looks very repetitive overall.

Example:

k141_5296188_14 ... Pfam PF00004 ATPase family associated with various cellular activities (AAA)
k141_2743843_13 ... TIGRFAM TIGR01027 proB: glutamate 5-kinase
k141_8488096_5 ... PANTHER PTHR23076 METALLOPROTEASE M41 FTSH

5) Pfam run separately For debugging, I also ran Pfam separately on the same predicted protein FASTA outside the pipeline, and it produced many diverse domain hits (not just PF00004). So Pfam itself appears to work correctly. Given that: InterProScan exits cleanly (exit 0) No fatal errors are reported Standalone Pfam behaves normally But the pipeline TSV output seems skewed / repetitive (only 37 hits in total) Could this be: An output parsing or aggregation issue? A filtering step in the pipeline? Something related to optional output binding? Or something still misconfigured in the InterProScan integration? Thanks again — really appreciate the help.

For reference, here is the

.command.sh

from the failing InterProScan task:

interproscan.sh \

--cpu 6 \

--input MEGAHIT-MetaBAT2Refined-BelfastCompost.40_cleaned.faa.fastq \

--applications PANTHER,ProSiteProfiles,ProSitePatterns,Pfam \

--disable-precalc --disable-residue-annot \

--enable-tsv-residue-annot --formats tsv \

--output-file-base MEGAHIT-MetaBAT2Refined-BelfastCompost.40_interproscan.faa

and in #C084HP11VPF this should be taken care of by default with:

withName: 'NFCORE_PROTEINANNOTATOR:PROTEINANNOTATOR:FUNCTIONAL_ANNOTATION:INTERPROSCAN' {
        publishDir = [
            path: { "${params.outdir}/functional_annotation/interproscan/${meta.id}/" },
            mode: params.publish_dir_mode,
            saveAs: { filename -> filename.equals('versions.yml') ? null : filename },
        ]
        ext.args = [
            "--applications ${params.interproscan_applications}",
            params.interproscan_enableprecalc ? '' : '--disable-precalc'
        ].join(' ').trim()
        cpus = 1 // bugs with more
    }

Thatswhere the NullPointerExpression comes from

https://github.com/nf-core/funcscan/blob/fa9db018e528ffb5149cdde928e2fa24e7c546fe/conf/modules.config#L129-L144 There?

Maybe that could influence it?

And @James Fellows Yates this is the command when i run it manually: and i just point it to the db where the sh command donwloaded with the db.