SpatialBenchVisium

This repository contains code used to perform analysis and create figures featured in our paper:

Du, Wang, Law, Amann-Zalcenstein et al. (2025) Benchmarking spatial transcriptomics technologies with the multi-sample SpatialBenchVisium dataset, Genome Biol 26:77.

Figure created with BioRender.

Data Availability

Our processed Visium and 10x scRNA-seq datasets, along with the code are available from zenodo: , data is also accessible through GEO: GSE254652.

Please cite our paper if you use our data and/or scripts in your research.

Index

Code to produce the reports are stored as Rmarkdown documents in analysis. Objects are saved in output.

Pre-processing

Spatial

Sample 709 FFPE CA: analysis/EDA_709_FFPE_CA.Rmd

Sample 713 FFPE CA: analysis/EDA_713_FFPE_CA.Rmd

scRNA-seq

analysis/sc_preprocessing.Rmd

Downstream analysis

Multi-sample feature selection, clustering, cell type deconvolution: analysis/FFPE_CA_multi-sample.Rmd

Pseudo-bulk differential expression analysis: as targets project under the targets_project folder

Running targets

Simply navigate to the targets_project folder from the zenodo tarball SpatialBenchVisium.tar.gz and run targets::tar_make() to run the entire pipeline, outputs are saved in the targets_project/output folder.

Figures

Figures 1-3, Supplementary figures S2-S7: analysis/figures.R

Figure 4 & 5, Supplementary figures S8-S11: run targets::tar_make() in the targets_project folder, figures are saved to targets_project/output