#'=========================================================================
#' apply_t_test
#' -------------------------------------------------------------------------
#' apply_t_test uses the statistical t_test to check if the fold-change of half
#' -life (HL) fragments and the fold-change intensity fragments respectively are
#' significant.
#'
#'
#' apply_t_test compares the mean of two neighboring fragments within the
#' same TU to check if the fold-change is significant.Fragments with distance
#' above threshold are not subjected to t-test.Dataframes with less than 3 rows
#' are excluded.
#'
#' The functions used are:
#'
#' 1. fragment_function: checks number of fragments inside TU, less
#' than 2 are excluded otherwise they are gathered for analysis.
#' 2. t_test_function: excludes dataframes with less than 3 rows,
#' makes fold-change and apply t-test, assign fragments names
#' and ratio, add columns with the corresponding p_values.
#'
#' @param inp SummarizedExperiment: the input data frame with correct format.
#' @param threshold integer: threshold.
#'
#' @return the SummarizedExperiment with the columns regarding statistics:
#' \describe{
#' \item{ID:}{The bin/probe specific ID.}
#' \item{position:}{The bin/probe specific position.}
#' \item{strand:}{The bin/probe specific strand.}
#' \item{intensity:}{The relative intensity at time point 0.}
#' \item{probe_TI:}{An internal value to determine which fitting model
#' is applied.}
#' \item{flag:}{Information on which fitting model is applied.}
#' \item{position_segment:}{The position based segment.}
#' \item{delay:}{The delay value of the bin/probe.}
#' \item{half_life:}{The half-life of the bin/probe.}
#' \item{TI_termination_factor:}{String, the factor of TI fragment.}
#' \item{delay_fragment:}{The delay fragment the bin belongs to.}
#' \item{velocity_fragment:}{The velocity value of the respective delay
#' fragment.}
#' \item{intercept:}{The vintercept of fit through the respective delay
#' fragment.}
#' \item{slope:}{The slope of the fit through the respective delay fragment.}
#' \item{HL_fragment:}{The half-life fragment the bin belongs to.}
#' \item{HL_mean_fragment:}{The mean half-life value of the respective
#' half-life fragment.}
#' \item{intensity_fragment:}{The intensity fragment the bin belongs to.}
#' \item{intensity_mean_fragment:}{The mean intensity value of the respective
#' intensity fragment.}
#' \item{TU:}{The overarching transcription unit.}
#' \item{TI_termination_fragment:}{The TI fragment the bin belongs to.}
#' \item{TI_mean_termination_factor:}{The mean termination factor of the
#' respective TI fragment.}
#' \item{seg_ID:}{The combined ID of the fragment.}
#' \item{pausing_site:}{presence of pausing site indicated by +/-.}
#' \item{iTSS_I:}{presence of iTSS_I indicated by +/-.}
#' \item{ps_ts_fragment:}{The fragments involved in pausing site or iTSS_I.}
#' \item{event_duration:}{Integer, the duration between two delay fragments.}
#' \item{event_ps_itss_p_value_Ttest:}{p_value of pausing site or iTSS_I.}
#' \item{p_value_slope:}{Integer, the p_value added to the inp.}
#' \item{delay_frg_slope:}{Integer, the slope value of the fit through the
#' respective delay fragment.}
#' \item{velocity_ratio:}{Integer, the ratio value of velocity from 2 delay
#' fragments.}
#' \item{event_position:}{Integer, position of the event added to the input.}
#' \item{FC_HL:}{Integer, the fold change value of 2 HL fragments.}
#' \item{FC_fragment_HL:}{String, the fragments corresponding to HL fold
#' change.}
#' \item{p_value_HL:}{Integer, the p_value added to the input of 2 HL
#' fragments.}
#' \item{FC_intensity:}{Integer, the fold change value of 2 intensity
#' fragments.}
#' \item{FC_fragment_intensity:}{String, the fragments corresponding to
#' intensity fold change.}
#' \item{p_value_intensity:}{Integer, the p_value added to the input of 2
#' intensity fragments.}
#' }
#'
#' @examples
#' data(stats_minimal)
#' apply_t_test(inp = stats_minimal, threshold = 300)
#'
#' @export
apply_t_test <- function(inp, threshold = 300) {
rowRanges(inp)$FC_fragment_HL <- NA
rowRanges(inp)$FC_HL <- NA
rowRanges(inp)$p_value_HL <- NA
rowRanges(inp)$FC_fragment_intensity <- NA
rowRanges(inp)$FC_intensity <- NA
rowRanges(inp)$p_value_intensity <- NA
uniqueTU <- unique(rowRanges(inp)$TU)
uniqueTU <- uniqueTU[grep("_NA|_T", uniqueTU, invert = TRUE)]
for (i in seq_along(uniqueTU)) {
# select ID, position, HL, HL fragments, intensity and intensity
# fragments for the corresponding TU
tu <-
as.data.frame(rowRanges(inp)[which(rowRanges(inp)$TU %in% uniqueTU[i]), c(
"ID",
"position",
"half_life",
"TU",
"HL_fragment",
"intensity",
"intensity_fragment",
"HL_mean_fragment",
"intensity_mean_fragment"
)])
# HL and intensity segments in the TU
hl_segs <-
tu[grep(paste0("\\Dc_\\d+", "$"),
tu$HL_fragment), "HL_fragment"]
int_segs <- tu[grep(paste0("\\I_\\d+", "$"),
tu$intensity_fragment), "intensity_fragment"]
hl_segs <- fragment_function(hl_segs)
int_segs <- fragment_function(int_segs)
# loop into all HL segments and apply t_test between consecutive segments
tryCatch({
inp <-
t_test_function(
data = inp,
seg = hl_segs,
param = "half_life",
o = "HL",
tu = tu,
threshold = threshold
)
}, error = function(e) {
})
if(length(int_segs) > 1){
inp <-
t_test_function(
data = inp,
seg = int_segs,
param = "intensity",
o = "intensity",
tu = tu,
threshold = threshold
)
}
}
return(inp)
}
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.