scRNA-seq 10x Flex Preprocessing

Cell Ranger multi setup for pooled 10x Flex libraries

What it does

This workflow covers the preprocessing step for pooled 10x Genomics Flex scRNA-seq libraries. It focuses on turning one multiplexed FASTQ set into per-sample filtered feature-barcode matrices using cellranger multi, along with the configuration files and OSC batch submission pattern needed to launch the run.

When to use it

Use this workflow when your wet-lab setup produced a 10x Flex experiment with multiple samples pooled into the same sequencing files. It is meant for the preprocessing stage only, before clustering or downstream single-cell analysis, and it is most useful when the main challenge is correctly mapping wet-lab sample metadata into the 10x multi configuration.

Prerequisites

Steps

Build the cellranger multi CSV from the wet-lab sample sheet

The README treats the configuration CSV as the key handoff between the wet-lab sample sheet and Cell Ranger. It maps the raw FASTQ identifier, FASTQ directory, reference, probe set, sample IDs, probe barcode IDs, and sample descriptions into one file for cellranger multi.

That mapping step is the core decision point in this workflow: once the CSV is correct, Cell Ranger can demultiplex pooled Flex data and write per-sample outputs without additional custom preprocessing code.

Pair the configuration CSV with an OSC batch script

The committed Slurm script is an OSC-style example that sets account, runtime, memory, working directory, and the cellranger multi command. In practice, the workflow is driven by two companion files:

  • the multi configuration CSV
  • a batch script that points cellranger multi at that CSV

This keeps the run command simple while making the metadata and demultiplexing settings explicit in the CSV.

Submit the Flex preprocessing job on OSC

#!/usr/bin/bash
#SBATCH --account PAS2505
#SBATCH --time=10:00:00
#SBATCH --nodes=1
#SBATCH --ntasks=8
#SBATCH --mem=64GB

CellRanger=/fs/ess/PCON0022/tools/cellranger-7.1.0/cellranger
cd /fs/ess/PAS2505/230801_Grayson_GSL-RH-3496/alignment/alignment_epi_data
${CellRanger} multi --id=Y12696_GraysonM_Naive-i_V1G_1 --csv=Y12696_GraysonM_Naive-i_V1G_1.csv --localcores=8 --localmem=64

The README also notes that the Slurm output file is the main place to monitor real-time run status after submission.

Review the per-sample summaries and collect filtered matrices

The documented outputs are per-sample web_summary.html reports and filtered feature-barcode matrices suitable for downstream Seurat or Scanpy workflows.

sample_filtered_feature_bc_matrix/
├── barcodes.tsv.gz
├── features.tsv.gz
└── matrix.mtx.gz

The practical handoff from this page is straightforward: once cellranger multi finishes, move from the web_summary.html reports into the per-sample filtered matrices for downstream QC, clustering, or format conversion in other workflows.

Gotchas / notes

  • The CSV content is the critical sample-demultiplexing input; mismatched probe barcode IDs or sample descriptions will propagate into the output.
  • The committed paths are OSC-specific examples and should be adapted to the actual project storage and account.
  • The README includes an externally hosted screenshot of the CSV layout, but there are no committed local figure assets in this workflow folder.
  • This page only covers pooled Flex preprocessing; downstream single-cell analysis happens in other workflow folders.

📄 View source on GitHub