Illumina RNA Pipeline¶

Paired-end Illumina RNA-seq pipeline tailored for metatranscriptomic data mapped against references from other danaSeq pipelines (assemblies from illumina_assembly/nanopore_assembly, or annotated MAGs from mag_analysis).

Quick Start¶

cd illumina_rna
./install.sh && ./install.sh --check

# Local (conda)
./run-illumina-rna.sh \
    --input /path/to/reads \
    --references /path/to/refs_dir \
    --outdir /path/to/output \
    --human_ref /path/to/bbtools_human_index \
    --sortmerna_refs /path/to/sortmerna_fastas

# Apptainer (HPC) -- auto-pulls SIF on first run
./run-illumina-rna.sh --apptainer --pull \
    --input /path/to/reads --references /path/to/refs --outdir /path/to/output

# SLURM profile
./run-illumina-rna.sh --input /path/to/reads --references /path/to/refs --outdir /path/to/output \
    -profile slurm --slurm_account def-myaccount \
    --conda_path ~/scratch/miniforge3/bin

Inputs¶

--input — directory of paired-end reads. Supports two naming conventions:
- Illumina: *_R1_*.fastq.gz / *_R2_*.fastq.gz
- BGI/MGI: *_1.fq.gz / *_2.fq.gz (and *.fastq.gz variants)
--references — directory of <name>.fasta files. If a sibling <name>.gff is present (typically the output of bakta/prokka annotation), gene-level counts are emitted via featureCounts; otherwise quantification is contig-level only via samtools idxstats.

Pipeline Stages¶

Step	Process	Tool	Description
1	`CLUMPIFY`	BBTools clumpify	Optical deduplication
2	`FILTER_BY_TILE`	BBTools filterbytile	Bad-tile filtering
3	`BBDUK_TRIM`	BBTools bbduk	Adapter trimming + length filter
4	`BBDUK_FILTER`	BBTools bbduk	Artifact + PhiX removal
5	`REMOVE_HUMAN`	BBTools removehuman	Host decontamination (optional)
6	`FASTQC`	FastQC	Per-sample QC report (optional)
7	`REMOVE_RRNA`	SortMeRNA	rRNA depletion (optional, on by default)
8	`BBMAP_INDEX`	BBMap	Reference index build
9	`MAP_READS_BBMAP`	BBMap + samtools	Per-(sample, reference) alignment
10	`FEATURECOUNTS`	subread featureCounts	Gene-level counts (when GFF provided)
11	`MERGE_GENE_COUNTS`	python/pandas	Genes × samples matrix per reference
12	`VIZ_PREPROCESS`	python	JSONs for the Svelte viz

Outputs¶

results/
├── preprocess/<sample>/                Cleaned reads + FastQC + SortMeRNA log
├── mapping/<sample>/<ref>/             sorted.bam(.bai), idxstats, flagstat, covstats
├── quantify/<sample>/<ref>/            featureCounts per-sample counts + summary
├── expression/<ref>/                   <ref>.gene_counts.tsv (merged, genes × samples)
├── viz/                                JSONs for the Svelte viz
└── pipeline_info/                      Nextflow timeline/report/trace/dag

Viz¶

cd viz
npm install
VIZ_DATA_DIR=/path/to/output/viz npm run dev
# open http://localhost:5173

Pages: Overview (sample × reference mapping-rate matrix, read-flow funnel), Samples (per-sample read flow + per-reference mapping detail), Expression (top-variance gene × sample heatmap with log₂ transform; per-reference dropdown), References (per-contig read totals).