Issue when rerunning the pipeline for MACS2 narrow peak

Description of the bug

The pipeline completed in default broad peak mode, and only required rerunning the MACS2 task in narrow peak mode.
But the process resumes from BWA alignment task which is unexpected and consumes a lot of time.

Command used and terminal output

$ nextflow run nf-core/atacseq -profile docker -resume -with-trace --narrow_peak --read_length 150 --input ./exp2_sample_sheet.csv --outdir ./results --fasta ./results/genome/Homo_sapiens.GRCh38.dna.primary_assembly.fa --gtf ./results/genome/Homo_sapiens.GRCh38.111._ensembl.gtf --bwa_index ./results/genome/index/bwa/

[5c/fa411f] process > NFCORE_ATACSEQ:ATACSEQ:PREPARE_GENOME:GTF2BED (Homo_sapiens.GRCh38.111._ensembl.gtf)                               [100%] 1 of 1 ✔
[14/5f793a] process > NFCORE_ATACSEQ:ATACSEQ:PREPARE_GENOME:TSS_EXTRACT                                                                  [100%] 1 of 1 ✔
[78/677e42] process > NFCORE_ATACSEQ:ATACSEQ:PREPARE_GENOME:CUSTOM_GETCHROMSIZES (Homo_sapiens.GRCh38.dna.primary_assembly.fa)           [100%] 1 of 1 ✔
[0d/b0fa56] process > NFCORE_ATACSEQ:ATACSEQ:PREPARE_GENOME:GET_AUTOSOMES (Homo_sapiens.GRCh38.dna.primary_assembly.fa.fai)              [100%] 1 of 1 ✔
[b5/8ee142] process > NFCORE_ATACSEQ:ATACSEQ:PREPARE_GENOME:GENOME_BLACKLIST_REGIONS (Homo_sapiens.GRCh38.dna.primary_assembly.fa.sizes) [100%] 1 of 1 ✔
[49/aeb1f4] process > NFCORE_ATACSEQ:ATACSEQ:PREPARE_GENOME:KHMER_UNIQUEKMERS (Homo_sapiens.GRCh38.dna.primary_assembly.fa)              [100%] 1 of 1 ✔
[c9/7a5968] process > NFCORE_ATACSEQ:ATACSEQ:INPUT_CHECK:SAMPLESHEET_CHECK (exp2_sample_sheet.csv)                                       [100%] 1 of 1, cached: 1 ✔
[61/e63750] process > NFCORE_ATACSEQ:ATACSEQ:FASTQ_FASTQC_UMITOOLS_TRIMGALORE:FASTQC (D6_MLN_REP2_T1)                                    [100%] 12 of 12, cached: 12 ✔
[e7/3a81d5] process > NFCORE_ATACSEQ:ATACSEQ:FASTQ_FASTQC_UMITOOLS_TRIMGALORE:TRIMGALORE (D3_DMSO_REP2_T1)                               [100%] 12 of 12, cached: 12 ✔
[11/6ac85d] process > NFCORE_ATACSEQ:ATACSEQ:FASTQ_ALIGN_BWA:BWA_MEM (D6_MLN_REP2_T1)                                                    [  0%] 0 of 12
[-        ] process > NFCORE_ATACSEQ:ATACSEQ:FASTQ_ALIGN_BWA:BAM_SORT_STATS_SAMTOOLS:SAMTOOLS_SORT                                       -
[-        ] process > NFCORE_ATACSEQ:ATACSEQ:FASTQ_ALIGN_BWA:BAM_SORT_STATS_SAMTOOLS:SAMTOOLS_INDEX                                      -
.
.
.
Other tasks follow

Relevant files

No response

System information

This seems as a better order to list the system information, from lowest to highest level.

Version of nf-core/atacseq: 2.1.2
Nextflow version: v24.10.3
Container engine: Docker 27.4.1
OS: Ubuntu 22.04.5 LTS x86_64
Executor: Local
Hardware: Desktop

Start by reproducing the supplied nf-core/atacseq 2.1.2 command with Docker and -resume, then inspect the trace and workflow caching around the BWA alignment and MACS2 narrow-peak tasks. The issue names no files or tests; done means rerunning with --narrow_peak reuses completed alignment work instead of restarting BWA unnecessarily.