permTestTx_customPick: Perform permutation test

View source: R/permTestTx_customPick.R

permTestTx_customPickR Documentation

Perform permutation test

Description

Perform permutation test for evaluating spatial association between a feature set and the customPick regions. The latter is defined by the customPick_function argument provided by users.

Usage

permTestTx_customPick(RS1 = NULL, txdb = NULL, type = "mature",
customPick_function = NULL, ntimes = 50, ev_function_1 = overlapCountsTx,
ev_function_2 = overlapCountsTx,  pval_z = FALSE, ...)

Arguments

RS1

The feature set to be randomized. It should be in the GRanges or GRangesList format.

txdb

A TxDb object.

type

A character object. Default is "mature". It accepts options "mature", "full", "fiveUTR", "CDS" or "threeUTR", with which one can get corresponding types of transcriptome regions.

customPick_function

A custom function needs to be inputted by users. The custom function should have two arguments: a TxDb object and a character object of transcript ids. It returns a part of region of each transcript.

ntimes

Randomization times.

ev_function_1

Evaluation function defines what statistic to be tested between RS1 and RS2. Default is overlapCountsTx.

ev_function_2

Evaluation function defines what statistic to be tested between each element in RSL and RS2. Default is overlapCountsTx.

pval_z

Boolean. Default is FALSE. If FALSE, the p-value is calculated based on the number of random evaluations is larger or less than the initial evaluation. If TRUE, the p-value is calculated based on a z-test.

...

Any additional parameters needed.

Details

Each feature in RS1 is only mapped with the customPick regions over its transcript (picked by the customPick_function). The output orig.ev is the number of features that have overlap with its customPick region. The set of randomized region sets is outputted as RSL. The overlapping counts between each set in RSL with RS2 is outputted as rand.ev.

Value

A list object, which is defined to be permTestTx.results class. It contains the following items:

  • RSL: Randomized region sets of RS1.

  • RS1: The feature set to be randomized.

  • RS2: The region set to be compared with the feature set.

  • orig.ev: The value of overlapping counts between RS1 and RS2.

  • rand.ev: The values of overlapping counts between each element in RSL and RS2.

  • pval: p-value of the test.

  • zscore: Standard score of the test.

See Also

plotPermResults

Examples

library(TxDb.Hsapiens.UCSC.hg19.knownGene)
txdb <- TxDb.Hsapiens.UCSC.hg19.knownGene
exons.tx0 <- exonsBy(txdb)
trans.ids <- sample(names(exons.tx0), 100)
RS1 <- randomizeTx(txdb, trans.ids, random_num = 100,
random_length = 200, type = 'CDS')
getCDS = function(txdb, trans.id){
cds.tx0 <- cdsBy(txdb, use.names=FALSE)
    cds.names <- as.character(intersect(names(cds.tx0), trans.id))
    cds = cds.tx0[cds.names]
    return(cds)
}

permTestTx_results <- permTestTx_customPick(RS1,txdb,
customPick_function = getCDS, ntimes = 5)

yue-wang-biomath/RgnTX documentation built on Aug. 24, 2023, 1:12 p.m.