qc_rna

Note

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

Quality control intended for metatranscriptomic (RNA-seq) reads. Mechanically the same pipeline as qc_dna — same modules, same defaults — but documented separately because the inputs and the appropriate --kneaddata_refdb choice differ.

What it does

Identical to qc_dna:

1. FastQC raw reads.
2. Trim Galore adapter/quality trimming.
3. Optional FASTX header fix (--fix_fastq_header).
4. KneadData decontamination using the references in --kneaddata_refdb.
5. FastQC post-QC reads.
6. Summary plot of read counts across stages.

The dispatcher in workflows/metagear.nf routes both qc_dna and qc_rna through the same QUALITY_CONTROL subworkflow (params.workflow.startsWith("qc_")). The difference is which references you pass:

• For DNA: a host genome (e.g. Homo_sapiens).
• For RNA: a host genome plus an rRNA reference (e.g. SILVA), so ribosomal reads are removed alongside host reads. Both go into the --kneaddata_refdb array; KneadData runs against each in turn.

If you don’t pass an rRNA reference, the workflow still runs but rRNA reads will dominate the output.

Inputs

• --input — CSV samplesheet (sample,fastq_1,fastq_2). See usage.md.
• --kneaddata_refdb — array of Bowtie2-indexed references. For metatranscriptomics, include at least one host genome and an rRNA reference.

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 + rRNA reference(s) for filtering.
--fix_fastq_header	boolean	false	Rewrite FASTQ headers via FASTX before KneadData.

Output

Same shape as qc_dna. Output paths are identical; only the MultiQC report and validated samplesheet are prefixed with qc_rna_:

Path (relative to --outdir)	Content
kneaddata/<sample>_paired_{1,2}.fastq.gz	Final clean reads — host- and rRNA-depleted.
kneaddata/<sample>_kneaddata.log	Per-sample log (one block per reference).
kneaddata/stats/qc_summary_plot.png	Visual breakdown of read losses by step.
fastqc/, trimgalore/	Intermediate FastQC and Trim Galore output.
pipeline_info/qc_rna_multiqc_report.html	Consolidated MultiQC report.

Expect the final read count to drop substantially compared with DNA — that’s the rRNA filtering working.

Example

nextflow run schirmer-lab/metagear -profile docker \
  --workflow qc_rna \
  --input raw_rna_samples.csv \
  --outdir qc_rna/ \
  --kneaddata_refdb '[/data/metagear/kneaddata/Homo_sapiens,/data/metagear/kneaddata/SILVA_128_LSURef]'

Notes

• Expect heavy attrition. A typical RNA library can be 70–90% rRNA; the cleaned output may be only a fraction of the raw input. This is by design.
• No mRNA enrichment is performed beyond rRNA depletion. If you have residual host mRNA you don’t want, add the host transcriptome to --kneaddata_refdb.
• Downstream compatibility. The output FASTQs are interchangeable with qc_dna output — you can feed them into microbial_profiles or gene_analysis. Just remember that abundance from RNA reflects activity, not presence.