Microbiome Mixture Analysis

Evaluate whether a microbiome sample is a mixture of two samples, by fitting a model for the number of read counts as a function of single nucleotide polymorphism (SNP) allele and the genotypes of two potential source samples. Lobo et al. (2021) .

R/mbmixture

R/mbmixture is an R package for evaluating whether a microbiome sample is the mixture of two source samples. We are thinking of shotgun sequencing data on the microbiome sample plus dense SNP genotype data on the two potential source samples. We assume that the data has been reduced to a three-dimensional array of read counts: the 3 possible SNP genotypes for the first sample × the 3 possible SNP genotypes of the second sample × the 2 possible SNP alleles on the reads.

We fit a model with contaminant probability p = proportion of the microbiome sample coming from the second sample and e = rate of sequencing errors.

Installation

You can install R/mbmixture from CRAN:

install.packages("mbmixture")

Alternatively, install it from R universe:

install.packages("mbmixture", repos=c("https://kbroman.r-universe.dev",
                                      "https://cloud.r-project.org"))

Or use remotes to install it from its GitHub source:

install.packages("remotes")
remotes::install_github("kbroman/mbmixture")

Vignette

A vignette describing the use of the package is available on the web. Or view it from within R by loading the package and then using the vignette() function.

library(mbmixture)
vignette("mbmixture", package="mbmixture")

Citation

To cite R/mbmixture in publications use:

Lobo AK, Traeger LL, Keller MP, Attie AD, Rey FE, Broman KW (2021) Identification of sample mix-ups and mixtures in microbiome data in Diversity Outbred mice. G3 (Bethesda) 11:jkab308 doi: 10.1093/g3journal/jkab308