Code
# Download the probe set
curl --output ../data/references/Visium_Human_Transcriptome_Probe_Set_v2.0_GRCh38-2020-A.csv \
https://cf.10xgenomics.com/supp/spatial-exp/probeset/Visium_Human_Transcriptome_Probe_Set_v2.0_GRCh38-2020-A.csvWe aligned all the raw Visium data using SpaceRanger (v3.0.1). We used the module installed on our HPC, but you can download it and find installation instructions here
We used the same GRCh38 reference genome described in CellRanger Alignment, along with the Visium Human Transcriptome Probe Set v2.0 required for CytAssist samples
The following shell script is submitted to a high-performance computing (HPC) cluster as a SLURM job array to process each of the Visium samples in parallel.
The script pulls a visium_samples_manifest.txt file from the data directory. It is a tab-separated file with no header, and the columns used by the script below (1-indexed) are:
| Column | Field | Example |
|---|---|---|
| 2 | sample name | sample_1 |
| 7 | fastq path | path/to/sample_1/fastq |
| 8 | image path | path/to/sample_1/image.tif |
| 9 | CytAssist image path | path/to/sample_1/cytaimage.tif |
| 10 | slide ID | V42A20-001 |
| 11 | capture area | A1 |
| 12 | Loupe manual alignment | path/to/sample_1/loupe_alignment.json |
#!/bin/bash
#SBATCH --account=
#SBATCH --job-name='spaceranger_%a'
#SBATCH --output=logs/spaceranger_%a.log
#SBATCH --partition=standard
#SBATCH --mem=128G
#SBATCH --cpus-per-task=16
#SBATCH --time=48:00:00
#SBATCH --array=1-14
mkdir -p logs
# Import modules
ml spaceranger/3.0.1
# Setting up variables
samples_manifest=../data/visium_samples_manifest.txt
ref=../data/references/refdata-gex-GRCh38-2020-A
probe_set=../data/references/Visium_Human_Transcriptome_Probe_Set_v2.0_GRCh38-2020-A.csv
# Setting up output dir
outputDir=../outputs/spaceranger_alignment
mkdir -p $outputDir
# Extracting sample for the current job
sample=$(cat $samples_manifest | cut -f2 | sed -n ${SLURM_ARRAY_TASK_ID}p)
fastq=$(cat $samples_manifest | cut -f7 | sed -n ${SLURM_ARRAY_TASK_ID}p)
slide=$(cat $samples_manifest | cut -f10 | sed -n ${SLURM_ARRAY_TASK_ID}p)
capture_area=$(cat $samples_manifest | cut -f11 | sed -n ${SLURM_ARRAY_TASK_ID}p)
image=$(cat $samples_manifest | cut -f8 | sed -n ${SLURM_ARRAY_TASK_ID}p)
cytaimage=$(cat $samples_manifest | cut -f9 | sed -n ${SLURM_ARRAY_TASK_ID}p)
loupe_alignment=$(cat $samples_manifest | cut -f12 | sed -n ${SLURM_ARRAY_TASK_ID}p)
echo "Analyzing Sample $sample"
# Running spaceranger
spaceranger count \
--id=$sample --sample=$sample \
--fastqs=$fastq \
--transcriptome=$ref --probe-set=$probe_set \
--slide=$slide --area=$capture_area \
--image=$image --cytaimage=$cytaimage \
--loupe-alignment=$loupe_alignment \
--localcores=16 --create-bam=false
mv $sample $outputDirDonor (n = 5), adjacent-normal-to-tumor (n = 2), and PDAC (n = 7) samples were profiled on the 10x Genomics Visium CytAssist platform using the 11 mm x 11 mm slide. Library preparation and sequencing were performed at the University of Michigan Advanced Genomics Core. The resulting feature-barcode matrices are loaded into Seurat as described in Spatial Transcriptomics Data Processing.