Count Regression for Correlated Observations with the Beta-binomial
corncob is an R package for modeling
relative abundance and testing hypotheses about the effect of covariates
on relative abundance. The corncob methodology was
specifically developed for modelling microbial abundances based on high
throughput sequencing data, such as 16S or whole-genome sequencing.
To install the corncob package, use the code below to
download the development version from Github.
# install.packages("remotes")
remotes::install_github("statdivlab/corncob")
library(corncob)Instead of installing corncob to your local system, you can use
corncob via the pre-compiled Docker image:
quay.io/fhcrc-microbiome/corncob.
The vignettes demonstrate example usage of all main functions (they
go over the same analysis, one with phyloseq objects and
one without phyloseq objects). Please file an issue if
you have a request for a tutorial that is not currently included. You
can see the vignette by using the following code:
library(corncob)
# Use this to view the vignette in the corncob HTML help
help(package = "corncob", help_type = "html")
# Use this to view the vignette as an isolated HTML file
utils::browseVignettes(package = "corncob")Note that R does not allow variable names to start with numbers.
Sometimes, when going directly from QIIME2 to phyloseq objects, taxa
names will be a large string starting with numbers. To clean these taxa
names for use with corncob, use
clean_taxa_names(my_phyloseq_object), see
?clean_taxa_names for details.
We additionally have a pkgdown website that contains
pre-built versions of our function documentation,
vignette,
and a version of the vignette
without phyloseq.
If you use corncob for your analysis, please cite our
manuscript:
Bryan D. Martin, Daniela Witten, and Amy D. Willis. (2020). Modeling microbial abundances and dysbiosis with beta-binomial regression. Annals of Applied Statistics, Volume 14, Number 1, pages 94-115.
An open-access preprint is available on arXiv here.
If you encounter a bug or would like make a change request, please file it as an issue here.