qc_dna

Note

Mirrored from the MetaGEAR Pipeline repository: docs/workflows/qc_dna.md (db08c99). Check there directly for up-to-date information.

Quality control for shotgun-metagenomic DNA reads. Produces decontaminated, adapter-trimmed paired reads that are ready for downstream analysis.

What it does

1. FastQC on the raw reads — pre-QC report.
2. Trim Galore for adapter and quality trimming. Process arg: --phred33 --quality 0 --stringency 5 --length 10 (light defaults; quality trimming is delegated to KneadData).
3. Optional FASTX header rewrite — only runs when --fix_fastq_header true. Use this if your reads have non-standard or Illumina-incompatible headers that confuse downstream tools.
4. KneadData for host decontamination. Process arg: --trimmomatic-options 'HEADCROP:15 SLIDINGWINDOW:4:15 MINLEN:50' --reorder --remove-intermediate-output --bypass-trf. Reads are aligned against the host genome(s) supplied via --kneaddata_refdb and host-matching pairs are dropped.
5. FastQC again on the decontaminated reads — post-QC report.
6. Summary plot of read counts across stages (raw → trimmed → host-removed → final).

qc_dna and qc_rna share this exact pipeline (the dispatcher in workflows/metagear.nf routes both via params.workflow.startsWith("qc_")). See qc_rna for the metatranscriptomic framing.

Inputs

• --input — CSV samplesheet (sample,fastq_1,fastq_2). See usage.md for the contract.
• --kneaddata_refdb — one or more Bowtie2-indexed host references for decontamination. Set this to the database installed by download_databases.

Parameters

Parameter	Type	Default	Controls
--input	path	(required)	Samplesheet of raw FASTQ pairs.
--outdir	path	(required)	Where to write results.
--kneaddata_refdb	array	[""]	Host genome(s) for read filtering (e.g. [/data/kneaddata/Homo_sapiens]).
--fix_fastq_header	boolean	false	Rewrite FASTQ headers via FASTX before KneadData. Enable if your reads come from a non-standard source and KneadData rejects them.

Output

Path (relative to --outdir)	Content
fastqc/<sample>_fastqc.{html,zip}	Pre-QC FastQC per raw FASTQ.
trimgalore/<sample>_*_val_{1,2}.fq.gz	Adapter-trimmed reads (intermediate).
trimgalore/<sample>_*_trimming_report.txt	Trim Galore log.
kneaddata/<sample>_paired_{1,2}.fastq.gz	Final clean reads. Use these as input to downstream workflows.
kneaddata/<sample>_kneaddata.log	Per-sample KneadData log.
kneaddata/stats/	Per-sample read-count CSV and the QC summary plot (qc_summary_plot.png).
fastqc/<sample>_clean_fastqc.{html,zip}	Post-QC FastQC per cleaned FASTQ.
pipeline_info/qc_dna_multiqc_report.html	Consolidated MultiQC report across all samples.
pipeline_info/qc_dna_samplesheet.valid.csv	Validated samplesheet as the pipeline saw it.

The “final clean reads” row is the only output you need to keep long-term; the rest are diagnostic.

Example

nextflow run schirmer-lab/metagear -profile docker \
  --workflow qc_dna \
  --input raw_samples.csv \
  --outdir qc/ \
  --kneaddata_refdb /data/metagear/kneaddata/Homo_sapiens

After this, build a clean.csv pointing at the kneaddata/<sample>_paired_{1,2}.fastq.gz files and feed it to microbial_profiles or gene_analysis.

Notes

• Trim Galore is intentionally light (--quality 0); the heavier quality trimming is done by KneadData’s internal Trimmomatic step with SLIDINGWINDOW:4:15 MINLEN:50. Don’t tighten Trim Galore unless you know why.
• HEADCROP:15 removes the first 15 bases of each read inside KneadData. This is appropriate for typical Illumina libraries with biased initial bases; if your library doesn’t have that artifact you may want to drop it via a custom config.
• Multiple host references. Pass an array to --kneaddata_refdb to filter against several genomes in one pass (e.g. human + mouse for xenograft samples).
• Single-end input is accepted — leave fastq_2 empty in the samplesheet.
• --fix_fastq_header is off by default. Turn it on only if KneadData errors out on header parsing.