Description Usage Arguments Value
Permutation-based differential abundance analysis
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 | 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" |
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.