`PICARD_COLLECTHSMETRICS` process is skipped due to empty input channel

Description of the bug

When the reference genome is not gzipped (i.e. genome.fa), PICARD_COLLECTHSMETRICS process is skipped due to the empty input channel here. As a result, only *.intervallist file is generated inside the enrichment_metrics folder, but the metrics files for individual samples are missing. When ch_gzi is removed here, it works.

Command used and terminal output

nextflow run nf-core/methylseq \
	-revision dev \
	-latest \
	-resume \
	-profile docker \
	--genome GRCh38 \
	--fasta genome.fa \
	--run_targeted_sequencing \
	--target_regions_file "$references/$probes" \
	--collecthsmetrics \
	--input "$output/samplesheet_methylseq.csv" \
	--outdir "$output/human/results" \
	-work-dir "$output/human/work"

Relevant files

System information

Version: 25.10.4 build 11173
Created: 10-02-2026 15:17 UTC (23:17 SGST)
System: Linux 5.15.0-157-generic
Runtime: Groovy 4.0.28 on OpenJDK 64-Bit Server VM 17.0.3-internal+0-adhoc..src
Encoding: UTF-8 (UTF-8)

Start in subworkflows/local/targeted_sequencing/main.nf around lines 78-91 and inspect how ch_gzi is passed to modules/nf-core/picard/collecthsmetrics/main.nf around line 15. Reproduce the command with an uncompressed genome.fa and compare the channel contents with the working case. Done means PICARD_COLLECTHSMETRICS runs and produces per-sample metrics in enrichment_metrics alongside the interval list.