moat() is the main entry point for MOAT. It validates a
SummarizedExperiment-like object, records the audit parameters, and returns
a stable moat_audit object. Design and correction diagnostics are
evaluated from metadata; batch diagnostics combine distance-based PERMANOVA,
PERMDISP, and PCoA audits. Leakage diagnostics evaluate repeated measures,
batch-driven validation leakage, and optional temporal leakage.
Usage
moat(
x,
outcome,
batch = NULL,
covariates = NULL,
subject = NULL,
time = NULL,
assay = "counts",
transform = "auto",
distances = c("aitchison", "bray"),
n_perm = 999,
feature_associations = TRUE,
verbose = TRUE
)Arguments
- x
A
SummarizedExperiment::SummarizedExperiment()object. Objects extendingSummarizedExperiment, such asTreeSummarizedExperiment, are accepted.- outcome
A single string naming the outcome variable in
colData(x).- batch
Optional character vector naming batch variables in
colData(x).- covariates
Optional character vector naming covariates in
colData(x).- subject
Optional single string naming the subject identifier variable in
colData(x).- time
Optional single string naming the time variable in
colData(x).- assay
A single string naming the assay to audit. Defaults to
"counts".- transform
A single string naming the microbiome transformation to use in distance-based audit modules. Use
"auto"to choose the default transformation for each distance. Defaults to"auto".- distances
A character vector naming microbiome distances to record for downstream audit modules. Defaults to
c("aitchison", "bray").- n_perm
A single positive integer with the planned number of permutations for downstream audit modules. Defaults to
999.- feature_associations
A single logical value indicating whether to screen individual features for batch associations. Defaults to
TRUE.- verbose
A single logical value indicating whether future audit modules should report progress. Defaults to
TRUE.
Examples
data("toy_moat")
audit <- moat(
toy_moat,
outcome = "outcome",
batch = "batch"
)
is_moat_audit(audit)
#> [1] TRUE