R/testing.R
combinatorial_association.Rd
Run differential association tests between between all combinations of a factor variable. Can be used as post-hoc test for regression.
combinatorial_association(ps, variable, tax = "genus", confounders = NULL, min_count = 10, in_samples = 0.1, independent_weighting = TRUE, standardize = TRUE, shrink = TRUE)
ps | A phyloseq object containing the taxa counts. |
---|---|
variable | The factor variable to permute. |
tax | The taxa level on which to run differential tests. Defaults to genus. |
confounders | A character vector containing the confounders that should be used. |
min_count | Minimum required number of average counts for a taxa. |
in_samples | Taxa must be present in at least this fraction of samples. |
independent_weighting | Whether to adjust p values by independent weighting or normal Benjamini-Hochberg. factors. |
standardize | Whether to standardize continuous variables to a mean of zero and a variance of 1. If True log fold changes for those variables denote are relative to a change of one standard deviation in the variable value. |
shrink | Whether to return shrunken log fold changes. Defaults to true. |
A data.table containing the results.
NULL#> NULL