ZicoSeq: Permutation-based differential abundance analysis
In chloelulu/ZicoSeq: Has not been decided

Description Usage Arguments Value

Permutation-based differential abundance analysis

ZicoSeq(
  meta.dat,
  comm,
  grp.name,
  adj.name = NULL,
  prev.filter = 0.1,
  abund.filter = 10,
  min.prop = 0,
  is.winsor = TRUE,
  winsor.qt = 0.97,
  is.prior = TRUE,
  prior.dist = c("BetaMix", "ZIBB"),
  post.method = c("sample", "mean"),
  post.sample.no = 25,
  link.func = list(function(x) x^0.25, function(x) x^0.5, function(x) x^0.75),
  link.d.func = list(function(x) 0.25 * x^(-0.75), function(x) 0.5 * x^(-0.5),
    function(x) 0.75 * x^(-0.25)),
  variance.EB = FALSE,
  df.prior = 10,
  perm.no = 99,
  strata = NULL,
  stats.combine.func = max,
  stage.no = 6,
  topK = NULL,
  stage.fdr = 0.75,
  stage.max.pct = 0.5,
  is.fwer = FALSE,
  is.tree.fdr = FALSE,
  tree = NULL,
  verbose = TRUE,
  return.comm = FALSE,
  return.perm.F = FALSE,
  ...
)

`meta.dat`	a data frame containing the sample information
`comm`	a matrix of counts, row - features (OTUs, genes, etc) , column - sample
`grp.name`	a character, variable of interest; it could be numeric or categorical; should be in "meta.dat"
`adj.name`	a character vector, variable(s) to be adjusted; they could be numeric or categorical; should be in "meta.dat"
`prev.filter`	features with prevalence (i.e., nonzero proportion) less than "prev.cutoff" or be filtered
`abund.filter`	features with a total counts less than "abund.cutoff" or be filtered
`min.prop`	Undetermined
`is.winsor`	a logical value indicating whether winsorization should be performed to replace outliers. The default is TRUE.
`winsor.qt`	the winsorization quantile, above which the counts will be replaced
`is.prior`	a logical value indicating whether to perform posterior inference based on some prior distribution on the proportion data
`prior.dist`	prior distribution, either two-component beta-binomial mixture ("BetaMix") or zeroinflated beta-binomial ("ZIBB")
`post.method`	method for posterior inference, either based on posterior sampling ("sample") or approximate posterior mean ("mean")
`post.sample.no`	the number of posterior samples if posterior sampling is used
`link.func`	a list of functions that connects the ratios to the covariates
`link.d.func`	a list of the derivative function of "link.func"; only need to specifiy when "post.method" is "mean"
`variance.EB`	a logical value indicating whehter to perform empirical Bayes based variance shrinkage
`df.prior`	the degree of freedom of the prior inverse gamma distribution for variance shrinkage
`perm.no`	the number of permutations; If the raw p values are of the major interest, set "perm.no" to at least 999
`strata`	a factor indicating the permutation strata; permutation will be confined to each stratum
`stats.combine.func`	function to combine the F-statistic for the omnibus test
`stage.no`	the number of stages if multiple-stage ratio stategy is used
`topK`	the number of dominant features that will be excluded in the initial stage ratio calculation
`stage.fdr`	the fdr cutoff below which the features will be excluded for calculating the ratio
`stage.max.pct`	the maximum percentage of features that will be excluded
`is.fwer`	a logical value indicating whether the family-wise error rate control (West-Young) should be performed
`is.tree.fdr`	a logical value indicating whether tree-based false discovery rate shuold be carried out
`tree`	a class of "phylo", the tree relats all the OTUs, and should have the same names in "comm"
`verbose`	a logical value indicating whether the trace information should be printed out
`return.comm`	a logical value indicating whether the wisorized, filtered "comm" matrix should be returned
`return.perm.F`	Undetermined
`...`	arguments passing to tree-based fdr control

A list with the elements

`call`	the call
`comm`	the wisorized, filtered "comm" matrix
`filter.ind`	a vector of logical values indicating which features are tested
`R2`	a matrix of percent explained variance (number of features by number of functions)
`F0`	a matrix of F-statistics (number of features by number of functions)
`RSS`	a matrix of residual sum squares (number of features by number of functions)
`df.model, df.residual`	degree of freedoms for the model and residual space
`p.raw`	the raw p-values based on permutations (not accurate if "perm.no" is small)
`p.adj.fdr`	permutation-based FDR-adjusted p-values
`p.adj.tree.fdr`	permutation-based tree FDR-adjusted p-values
`p.adj.fwer`	permutation-based FWER-adjusted (West-Young) p-values
`tree.fdr.obj`	the object returned by the "TreeFDR"