Repository files navigation

The goal of this package is to identify response to a stimulant inCyTOF/Flow cytometry stimulation assays by labeling cells as respondedor not based on an unsupervised high dimensional approach. Starting fromthe annotated cell populations either through automated clustering suchas FlowSOM or traditional cell gating, the primary functionHDStIM()follows a heuristic approach to label cells as responding ornon-responding.

For a combination of cell population and stimulation type (e.g., CD127+T-helper cells and interferon-alpha),HDStIM() starts by performingk-means clustering on the combined set of cells from stimulated andunstimulated samples. K-means clustering is performed on expression dataof all the state markers combined. Upon clustering using a contingencytable, a Fisher’s exact test determines the effect size and thestatistical significance of partitioning. Cells form the combinationsthat pass the Fisher’s exact test are labelled as responding.

You can install the released version of stimcellselector fromCRAN with:

install.packages("HDStIM")

And the development version fromGitHub with:

# install.packages("devtools")devtools::install_github("niaid/HDStIM")

Rohit Farmer:rohit.farmer@nih.gov,rohit.farmer@gmail.com