nf-reprocessing-public-10x

Nextflow pipeline for loading and reprocessing public 10x datasets from GEO, SRA, ENA, or ArrayExpress.

Repo structure

Path	Description
main.nf	Main Nextflow pipeline entry point
nextflow.config	Pipeline configuration — LSF executor, Singularity, and default params
workflow/main.nf	Core workflow: metadata fetch, download, and STARsolo alignment
subworkflows/	Download and STARsolo alignment subworkflows
modules/	Individual process modules
examples/datasets.tsv	Example input file
examples/RESUME	Example run script

Usage

nextflow run main.nf --datasets <datasets.tsv> [OPTIONS]

Parameters

Parameter	Description	Default
--datasets	Path to a TSV file with dataset and sample IDs	required
--outdir	Directory to save results	results
--default_specie	Species to assign when metadata is missing or unknown (human or mouse). Samples without a resolved species are skipped by STARsolo.	null
--no_infer_specie	Skip reading species from metadata; assign --default_specie to all samples. Requires --default_specie.	false
--metaonly	Only fetch metadata — skip downloading and alignment	false
--help	Print help message and exit	—

Input file format

The --datasets TSV must have a header row with two columns:

dataset_id	sample_id
GSE230685	GSM7232572,GSM7232573
E-MTAB-9221	ERS4689152,ERS4689153
PRJEB37166	ERS4605100,ERS4605101

• dataset_id — GEO series (GSE*), BioProject (PRJEB*/PRJNA*), or ArrayExpress accession (E-MTAB-*)
• sample_id — comma-separated list of sample accessions belonging to that dataset

See examples/datasets.tsv for a full example.

Quick example

nextflow run main.nf \
  --datasets examples/datasets.tsv \
  --outdir results \
  --default_specie human \
  -resume

Output structure

results/
├── raw/
│   └── <dataset_id>/
│       ├── fastq/
│       │   └── <sample_id>/        FASTQs (R1, R2, I1)
│       ├── bam/
│       │   └── <sample_id>/        10x BAM files
│       └── sra/
│           └── <sample_id>/        SRA files
├── starsolo/
│   └── <dataset_id>/               STARsolo count matrices and QC stats
├── metadata/
│   └── <dataset_id>/               Metadata files (links, parsed TSVs, SOFT, etc.)
├── index/
│   ├── fastq.csv                   Index of all published FASTQs
│   ├── bam.csv                     Index of all published BAMs
│   ├── sra.csv                     Index of all published SRA files
│   └── starsolo.csv                Index of all STARsolo outputs
├── versions.yml                    Software versions used by each process
└── mapping_qc_stats.tsv            Per-sample STARsolo mapping QC statistics

Species handling

By default (--no_infer_specie not set), species is read from sample metadata. If it is blank or unrecognised:

• If --default_specie is set, that species is used and a warning is logged.
• If --default_specie is not set, the sample is assigned UNKNOWN and skipped by STARsolo.

Use --no_infer_specie to bypass metadata entirely and force all samples to --default_specie.

Requirements

• Nextflow >=26.04.1
• Singularity (or Docker for local runs)
• LSF cluster (or adjust nextflow.config executor for local use)