One function for all taxonomic feature statistical tests.
Usage
test_taxa(
data.obj,
test.type = c("difference", "trend", "volatility", "association"),
subject.var = NULL,
time.var = NULL,
group.var = NULL,
adj.vars = NULL,
time.points = NULL,
feature.level = "original",
feature.dat.type = c("count", "proportion", "other"),
prev.filter = 0.1,
abund.filter = 1e-04,
feature.mt.method = c("fdr", "bonferroni", "none"),
feature.sig.level = 0.05,
...
)Arguments
- data.obj
A MicrobiomeStat data object containing feature.tab, meta.dat, and optionally feature.ann, tree, and feature.agg.list.
- test.type
Type of test: - "difference": Test for group differences (cross-sectional or paired) - "trend": Test for temporal trends (longitudinal) - "volatility": Test for temporal instability (longitudinal) - "association": Test for associations with continuous variables
- subject.var
Character. Name of the subject/sample ID variable in meta.dat. Required for longitudinal and paired designs.
- time.var
Character. Name of the time variable in meta.dat. NULL for cross-sectional studies.
- group.var
Character. Name of the grouping variable (e.g., treatment, condition).
- adj.vars
Character vector. Names of adjustment/covariate variables.
- time.points
Time point specification. Can be: - NULL: use all available time points (auto-detect design) - Single value: cross-sectional at that time point - Vector of 2: paired design (baseline, followup) - Vector of >2: longitudinal with specific time points - List: list(baseline = "T0", followup = c("T1", "T2"))
- feature.level
Character vector. Taxonomic level(s) for aggregation. Options include column names from feature.ann or "original" for no aggregation.
- feature.dat.type
Data type: "count", "proportion", or "other".
- prev.filter
Numeric. Minimum prevalence threshold (0-1) for filtering.
- abund.filter
Numeric. Minimum abundance threshold for filtering.
- feature.mt.method
Multiple testing correction: "fdr", "bonferroni", or "none".
- feature.sig.level
Significance threshold (default 0.05).
- ...
Additional arguments passed to the underlying function.
Examples
if (FALSE) { # \dontrun{
data(peerj32.obj)
# Cross-sectional group comparison
test_taxa(peerj32.obj, "difference",
group.var = "group",
feature.level = "Genus")
# Longitudinal trend test
test_taxa(peerj32.obj, "trend",
subject.var = "subject",
time.var = "time",
group.var = "group",
feature.level = "Genus")
} # }