SCRAT: Single Cell R Analysis Toolkit
Exploring single cell transcriptome heterogeneity with self-organizing maps
We have developed a suite of Single Cell R-Analysis Tools (SCRAT) based on self-organizing maps (SOMs) to deconstruct cellular heterogeneity, reconstruct lineage relationships, and perform gene set enrichment analysis using single-cell transcriptomes. A SOM displays similarity relationships in a 2D heat map in which each unit represents a set of genes (metagenes) whose expression patterns correlate across single cells. Each metagene’s spatial proximity reflects similarity across metagenes, with metagenes located in the same fixed positions across all SOM visualizations and groups of metagenes reflect cellular signatures. Therefore, each cell can be viewed as an intuitive and easy-to-interpret "portrait" of transcriptional activity. We have applied these tools to analyze transcriptome states emerging during differentiation of induced pluripotent stem cells into hepatocyte-like cells, and self-assembly of human liver bud organoids. This analytic framework will be broadly useful for exploring cell heterogeneity in complex tissues, during directed differentiation, or in drug response.
Given an input data matrix containing expression data for single cells, the SCRAT pipeline will generate a summary report (Summary.html) where all of the data analyses can be accessed and browsed by the user. In the vignette, we have described most of the features and individual commands that will allow the user to customizes data input and output.
Please contact us with any questions or suggestions!
Gray Camp: firstname.lastname@example.org
Henry Loeffler-Wirth: email@example.com