Skip to contents

disize

The existing methods for RNAseq normalization are DESeq2’s median of ratios (MoR) and edgeR’s trimmed mean of M values (TMM) . These methods however do not include information about the experimental design when trying to estimate size factors, and can fail for more complex study designs.

Design informed size factor estimation (or disize) is an alternative normalization method that jointly models gene expression and batch-effects following a specified design to gain precision on size factor estimates.

Usage

Take a look at the Get started page to familiarize yourself with disize.

Installation

As disize is not yet on CRAN, installation is not a one-liner with install.packages:

With remotes 

# Install disize
remotes::install_github("https://github.com/toddmccready/disize")

# Set up CmdStan toolchain
cmdstanr::install_cmdstan()

With rv

Add the following entry to your rproject.toml file (if not already present):

dependencies = [
    # ...
    { name = "disize", git = "https://github.com/toddmccready/disize", branch = "main" },
    # ...
]

Then install the CmdStan toolchain in R:

cmdstanr::install_cmdstan()