An introduction to the MoleculeExperiment Class

Bárbara Zita Peters Couto1,2,3, Nick Robertson1,2,3, Mu-Wei Chung1,2,3, Ellis Patrick1,2,3,4 and Shila Ghazanfar1,2,3

1School of Mathematics and Statistics, The University of Sydney, Sydney, NSW, 2006, Australia
2Charles Perkins Centre, The University of Sydney, Sydney, NSW, 2006, Australia
3Sydney Precision Data Science Centre, The University of Sydney, Sydney, NSW, 2006, Australia
4Westmead Institute for Medical Research, University of Sydney, Australia

Contents

1 MoleculeExperiment

The R package MoleculeExperiment contains functions to create and work with objects from the new MoleculeExperiment class. We introduce this class for analysing molecule-based spatial transcriptomics data (e.g., Xenium by 10X, CosMx SMI by Nanostring, and Merscope by Vizgen, among others).

1.1 Why the MoleculeExperiment class?

The goal of the MoleculeExperiment class is to: 1. Enable analysis of spatial transcriptomics (ST) data at the molecule level, independent of aggregating to the cell or tissue level. 2. Standardise molecule-based ST data across vendors, to hopefully facilitate comparison of different data sources and common analytical and visualisation workflows. 3. Enable aggregation to a SpatialExperiment object given combinations of molecules and segmentation boundaries.

1.2 Installation

The latest release of MoleculeExperiment can be installed using:

if (!require("BiocManager", quietly = TRUE)) {
  install.packages("BiocManager")
}
BiocManager::install("MoleculeExperiment")

2 Minimal example

1. Load required libraries.

library(MoleculeExperiment)
library(ggplot2)
library(EBImage)

2. Create MoleculeExperiment object with example Xenium data, taken over a small patch.

repoDir <- system.file("extdata", package = "MoleculeExperiment")
repoDir <- paste0(repoDir, "/xenium_V1_FF_Mouse_Brain")

me <- readXenium(repoDir, keepCols = "essential")
me
#> MoleculeExperiment class
#> 
#> molecules slot (1): detected
#> - detected:
#> samples (2): sample1 sample2
#> -- sample1:
#> ---- features (137): 2010300C02Rik Acsbg1 ... Zfp536 Zfpm2
#> ---- molecules (962)
#> ---- location range: [4900,4919.98] x [6400.02,6420]
#> -- sample2:
#> ---- features (143): 2010300C02Rik Acsbg1 ... Zfp536 Zfpm2
#> ---- molecules (777)
#> ---- location range: [4900.01,4919.98] x [6400.16,6419.97]
#> 
#> 
#> boundaries slot (1): cell
#> - cell:
#> samples (2): sample1 sample2
#> -- sample1:
#> ---- segments (5): 67500 67512 67515 67521 67527
#> -- sample2:
#> ---- segments (9): 65043 65044 ... 65070 65071

3. Use standardised data in ME object for molecule-level analyses. For example, plot a simple digital in-situ, with cell boundaries overlaid.

ggplot_me() +
  geom_polygon_me(me, assayName = "cell", fill = "grey") +
  geom_point_me(me) +
  # zoom in to selected patch area
  coord_cartesian(
    xlim = c(4900, 4919.98),
    ylim = c(6400.02, 6420)
  )