crumblr
Count ratio uncertainty modeling base linear regression
Bioconductor version: 3.23 · Package version: 1.4.5
Crumblr enables analysis of count ratio data using precision weighted linear (mixed) models. It uses an asymptotic normal approximation of the variance following the centered log ration transform (CLR) that is widely used in compositional data analysis. Crumblr provides a fast, flexible alternative to GLMs and GLMM's while retaining high power and controlling the false positive rate.
Installation
if (!require("BiocManager", quietly = TRUE))
install.packages("BiocManager")
BiocManager::install("crumblr") Details
| Maintainer | Gabriel Hoffman <gabriel.hoffman@mssm.edu> |
| Author | Gabriel Hoffman [aut, cre] (ORCID: <https://orcid.org/0000-0002-0957-0224>) |
| License | Artistic-2.0 |
| URL | https://DiseaseNeurogenomics.github.io/crumblr |
| Bug Reports | https://github.com/DiseaseNeurogenomics/crumblr/issues |
| Downloads rank | 154 |
| Source branch | RELEASE_3_23 |
| biocViews | BatchEffect, Clustering, DifferentialExpression, DimensionReduction, Epigenetics, FunctionalGenomics, GeneExpression, Normalization, Preprocessing, QualityControl, RNASeq, Regression, SingleCell, Software, Transcriptomics |
Documentation
- Using crumblr in practice
- Normal approximation vs. empirical simulation
- Integration with dreamlet / SingleCellExperiment
Download
Dependencies
Depends: R (>= 4.4.0), ggplot2, methods
Imports: Rdpack, viridis, tidytree, variancePartition (>= 1.36.3), SingleCellExperiment, ggtree, dplyr, stats, MASS, Rfast, dirmult
Suggests: BiocStyle, RUnit, knitr, rmarkdown, dreamlet, muscat, ExperimentHub, scater, reshape2, glue, tidyverse, BiocGenerics, compositions