Skip to content

Latest commit

 

History

History
97 lines (66 loc) · 4.31 KB

README.md

File metadata and controls

97 lines (66 loc) · 4.31 KB
Raw

Tweedieverse: Differential analysis of omics data based on the Tweedie distribution

Himel Mallick, Ali Rahnavard 2022-05-03

Introduction

Tweedieverse is an R package for differential analysis of omics data implementing a range of statistical methodology based on the Tweedie distribution.

Unlike traditional single-omics tools, Tweedieverse is technology-agnostic and can be applied to both count and continuous measurements arising from diverse high-throughput technologies (e.g., transcript abundances from bulk and single-cell RNA-Seq studies in the form of UMI counts or non-UMI counts, microbiome taxonomic and functional profiles in the form of counts or relative abundances, and compound abundance levels or peak intensities from metabolomics and other mass spectrometry-based experiments, among others).

The software includes multiple analysis methods (e.g., self-adaptive, zero-inflated, and non-zero-inflated statistical models) as well as multiple customization options such as the inclusion of random effects and multiple covariates along with several data exploration capabilities and visualization modules in a unified estimation umbrella.

Installation

To install the latest release version of Tweedieverse from CRAN (not yet available) run the following command:

install.packages("Tweedieverse")
library(Tweedieverse)

Alternatively, the latest development version of Tweedieverse can be loaded using the following command (execute from within a fresh R session):

install.packages('devtools')
library(devtools)
devtools::install_github("himelmallick/Tweedieverse")
library(Tweedieverse)

After installing Tweedieverse, please make sure the following package versions are also installed (a prerequisite for zero-inflated Tweedie models):

devtools::install_version("statmod", version = "1.4.33", repos ="http://cran.us.r-project.org")
devtools::install_version("cplm", version = "0.7-8", repos = "http://cran.us.r-project.org")

Basic Usage

Tweedieverse(features, metadata, output)

Input

Tweedieverse requires two input files:

  • features: A data frame of omics features such as taxa, genes, transcripts, metabolites, etc.
  • metadata: A data frame of metadata to be associated.

For full options, check out the user manual or type ?Tweedieverse in your R console.

Output

A data frame containing coefficient estimates, p-values, and q-values (multiplicity-adjusted p-values) are returned, along with other parameter estimates from the fitted per-feature models.

Getting Started with Tweedieverse

Check out the Tweedie Labs repository for a collection of walkthrough tutorials (available as source codes, cloud-compatible images, and installable packages) on how to use Tweedieverse with various omics data types.

Citation

To cite Tweedieverse in publications, please use:

Mallick, H, Chatterjee, S, Chowdhury, S, Chatterjee, S, Rahnavard, A, Hicks, SC. Differential expression of single-cell RNA-seq data using Tweedie models. Statistics in Medicine. 2022; 41( 18): 3492- 3510. doi:10.1002/sim.9430

To cite the Tweedieverse software, please use:

Mallick H et al. (2021). Tweedieverse - A Unified Statistical Framework for Differential Analysis of Multi-omics Data. R package, https://github.com/himelmallick/Tweedieverse.

Issues

We are happy to troubleshoot any issues with the package. Please contact the maintainer via email or open an issue in the GitHub repository.