README.md

In lulizou/spASE: Detecting Allele-Specific Expression in Spatial Transcriptomics

spASE

• Installation
• Quick start
• Hypothesis testing
• Visualization
• Reproducibility
• Citation

R package for detecting 2D spatial patterns of allele-specific expression in spatial transcriptomics data while controlling for cell type. Provides functions for estimating, plotting, and testing for significant ASE patterns. This repository also contains the code to reproduce the results in the manuscript.

Installation

devtools::install_github("lulizou/spASE")

Quick start

The main input to spASE includes:

• maternal counts matrix (rows are genes, columns are spots or cells)
• paternal counts matrix
• spatial coordinates.

Note that spASE currently modifies many functions from the spacexr package, so it is always good to load them in order so that the spASE version is used:

library(spacexr)
library(spASE)

Hypothesis testing

Testing for…

1. Overall bias (i.e. pseudo-bulk):

bias_results <- scase(
  matrix1 = maternal_counts_matrix,
  matrix2 = paternal_counts_matrix,
  min.cells = 100, # set to something reasonable
  cores = 1, # can add more
  verbose = T # to enable progress bar
)

Note that this can also be run on a subset of genes by specifying them as a string vector using the genes = c(...) argument.

1. Overall spatial pattern (i.e. no cell type effect):

overall_spatial_results <- spase(
  matrix1 = maternal_counts_matrix,
  matrix2 = paternal_counts_matrix,
  covariates = coords,
  min.pixels = 100, # set to something reasonable
  cores = 1, # can add more
  verbose = T # to enable progress bar
)

Note that coords must have the first column as the pixel ID (matching the column names of the count matrices) and the second two columns as the spatial coordinates.

Visualization

Once you have overall_spatial_results, we can visualize using the plotSpase function as follows:

plotSpase(
  matrix1 = maternal_counts_matrix,
  matrix2 = paternal_counts_matrix,
  covariates = coords,
  spasefit = overall_spatial_results,
  coords = NULL, # to downsample could set e.g coords = coords |> sample_n(1e4)
)

Reproducibility

The data is available in the Broad single cell portal at SCP1692.

To reproduce the results in the manuscript, see the analysis folder. To see the knitted Quarto documents, see:

• 00_preprocess.md - details on the alignment and custom scripts for running RCTD, C-SIDE, and spASE
• 01_cell_type_maps.md - cell type plots for each sample
• 02_compare_visium_slideseq_cerebellum.md - comparison of the Visium and Slide-seqV2 on different cerebellum samples
• 03_simulations.md - simulation results
• 04_overall_results_summary_figures.md - main summary figures and tables from running spASE
• 05_celltype_xchr_figures.md - main figure comparing cell type-driven and within cell type spatial ASE
• 06_visium_slideseq_mixtures_overdispersion.md - supplemental figures showing count distributions of perfect binomial sampling, spatial transcriptomics, and Smart-seq3 data
• 07_129_CAST_analysis.md - supplemental figures on comparing CAST and 129 transcripts
• 08_xchr_inactivation.md - supplemental figures showing X-chromosome genes for each sample

Citation

The (old) pre-print is here.

lulizou/spASE documentation built on May 22, 2024, 5:24 a.m.