selNode: Select branches meeting certain criteria

In fionarhuang/treeclimbR: An algorithm to find optimal signal levels in a tree

Select branches meeting certain criteria

Description

Select branches in a tree meeting the specified criteria in terms of number of leaves and the count proportion. Note that only internal branch nodes are considered - no individual leaves will be returned.

Usage

selNode(
  pr = NULL,
  obj = NULL,
  assay = 1,
  data = NULL,
  tree = NULL,
  minTip = 0,
  maxTip = Inf,
  minPr = 0,
  maxPr = 1,
  skip = NULL,
  all = FALSE
)

Arguments

pr	A named numeric vector to provide proportions of entities. If this is provided, obj and data will be ignored.
obj	A TreeSummarizedExperiment object. Only used if pr is NULL.
assay	The index or name of the assay of obj to use for estimating node count proportions. Only used if obj is not NULL.
data	Either a count table with entities in rows and samples in columns, or a list with pi and theta estimates (the output of parEstimate). Only used if pr and obj are NULL.
tree	A phylo object. If obj is used as input, the tree will be extracted from the rowTree of obj.
minTip	the minimum number of leaves in the selected branch.
maxTip	The maximum number of leaves in the selected branch.
minPr	The minimum count proportion of the selected branch in a sample. A value between 0 and 1.
maxPr	The maximum count proportion of the selected branch in a sample. A value between 0 and 1.
skip	A character vector of node labels. These nodes can not be descendants or the ancestors of the selected branch.
all	A logical scalar. If FALSE (default), the branch node of a single branch, which meets the requirements and has the minimum count proportion of branches meeting the requirements, is returned; otherwise branch nodes of all branches meeting the requirements are returned.

Value

A data.frame with node information for the selected internal node(s).

Author(s)

Ruizhu Huang, Charlotte Soneson

Examples

suppressPackageStartupMessages({
    library(TreeSummarizedExperiment)
})

## Generate example data
set.seed(1)
data(tinyTree)
toyTable <- matrix(rnbinom(40, size = 1, mu = 10), nrow = 10)
colnames(toyTable) <- paste(rep(LETTERS[seq_len(2)], each = 2),
                            rep(seq_len(2), 2), sep = "_")
rownames(toyTable) <- tinyTree$tip.label

## Estimate entity proportions from count matrix under a Dirichlet
## Multinomial framework, and use this as the input for selNode
dat <- parEstimate(obj = toyTable)
selNode(tree = tinyTree, data = dat, all = TRUE)
selNode(tree = tinyTree, data = dat,
        minTip = 4, maxTip = 9, minPr = 0, maxPr = 0.8, all = TRUE)

## Alternatively, directly provide the proportions vector
selNode(tree = tinyTree, pr = dat$pi, all = TRUE)

## Return only branch with lowest proportion among valid ones
selNode(tree = tinyTree, pr = dat$pi, all = FALSE)

## Start instead from a TreeSummarizedExperiment object
lse <- TreeSummarizedExperiment(rowTree = tinyTree,
                                assays = list(counts = toyTable))
selNode(obj = lse, assay = "counts", all = TRUE)

## Don't allow node 1 to be included
selNode(obj = lse, assay = "counts", skip = 1, all = TRUE)

fionarhuang/treeclimbR documentation built on Nov. 7, 2024, 4:17 a.m.