Skip to content

Latest commit

 

History

History
56 lines (39 loc) · 4.28 KB

README.md

File metadata and controls

56 lines (39 loc) · 4.28 KB
title author data output vignette
Vignettes
Dylan Cable
December 22, 2021
html_document pdf_document rmarkdown::html_vignette
keep_md
true
default
keep_md
true
%\VignetteIndexEntry{vignette-readme} %\VignetteEncoding{UTF-8} %\VignetteEngine{knitr::rmarkdown}

Vignettes: cell type identification and cell type-specific differential expression in spatial transcriptomics

Here, we will present examples and tutorials for how to run our computational methods for cell type identification (RCTD) and differential expression (C-SIDE) on spatial transcriptomics datasets. You may access RCTD and C-SIDE within our open-source R package here.

In total, we currently have 9 vignettes demonstrating the various applications of our software. These vignettes will be organized and referenced below, including applications to Slide-seq, Visium, and MERFISH data. In general, you can either view the html files (as linked to here) to view the vignette output, or you can run the raw R-markdown files on your own machine.

Cell type identification with RCTD

The best vignette for getting started with RCTD is spatial transcriptomics vignette.

RCTD can assign single cell types or cell type mixtures to spatial transcriptomics spots. RCTD has three modes: doublet mode, which assigns 1-2 cell types per spot and is recommended for technologies with high spatial resolution such as Slide-seq and MERFISH; full mode, which assigns any number of cell types per spot and is recommended for technologies with poor spatial resolution such as 100-micron resolution Visium; multi mode, an extension of doublet mode that can discover more than two cell types per spot as an alternative option to full mode. We demonstrate each mode in the following figures:

Cell type-specific differential expression with C-SIDE

The best vignette for getting started with C-SIDE is differential expression vignette.

C-SIDE can detect differential expression (DE) along one or multiple user-defined axes, termed explanatory variables. Although the possibilities are not limited to what is presented here, we present the following examples:

Batch processing of multiple experimental replicates + population-level DE inference

Finally, when multiple experimental replicates are available, RCTD and C-SIDE can be run in batch across replicates as shown in Population-level RCTD and C-SIDE. This approach also allows for population-level differential expression statistical inference.