logLin

OTU counts table. Rows: samples, columns: OTUs.

data

numbers of colums containing OTU counts.

count.columns

count threshold to transition from linear to log transform.

With 'zero.na=TRUE', the original zero count datapoints become 'NA' 
(useful for analysis of pairwise correlations between OTUs).

zero.na

Normalizes counts based on total counts per sample, then uses log-linear-hybrid
transformation on the data.

Poisson-lognormal generalized linear mixed model analysis of multivariate counts data using MCMC, aiming to infer the changes in relative proportions of individual variables. The package was originally designed for sequence-based analysis of microbial communities ("metabarcoding", variables = operational taxonomic units, OTUs), but can be used for other types of multivariate counts, such as in ecological applications (variables = species). The results are summarized and plotted using 'ggplot2' functions. Includes functions to remove sample and variable outliers and reformat counts into normalized log-transformed values for correlation and principal component/coordinate analysis. Walkthrough and examples: http://www.bio.utexas.edu/research/matz_lab/matzlab/Methods_files/walkthroughExample_mcmcOTU_R.txt.

logLin: prepares OTU counts data for PCA analysis using log-linear-hybrid transform

Description

Usage

Arguments

Value

References