SigWin {SigWinR} | R Documentation |
Calculate significant local changes in a sequence of data. The function tests the input sequence against the null hypothesis that the value in the sequence does not depend on position. The function produces p-values for the probability that the sequence is locally higher or lower than average for all possible positions for a given window size. The p-value is corrected for multiple testing using a method of choice.
Two variants of SigWin are implemented: one for simple ordered sequences without location information , and a second for a sequence of localized samples (for example expression measurements at known positions on a chromosome).
SigWin(s,window.size,pos=NULL,antiRidge=FALSE,adjust.method="BH",circular=FALSE,resolution=400)
s |
sequence of numeric data |
window.size |
size of window to use. For non localized (pos=NULL ) sequences the
window size is in sequence elements. For localized data the winow size is relative
to the location values given in the pos parameter |
pos |
optional sequence of sample location data |
antiRidge |
look for positive (antiRidge=FALSE) or negative deviations (antiRidge=TRUE). |
adjust.method |
method for multiple testing correction (see p.adjust for valid options). |
circular |
Assume the sequence is circular. If set to TRUE the window can extend
over the edges of the sequence into the other ende of the sequence. |
resolution |
resolution of SigWin calculation for localized sequences. Value is ignored for calculations without location data |
A sequence containing the p-values corrected for multiple testing. For simple non localized sequences the length of the output sequence is the length of the input sequence plus one minus the window size. For localized sequences the length of the output is equal to the resolution parameter.
This function can be used for purposes where Rigeogram is not needed or wanted such as long sequences. To generate a ridgeogram the Ridgeogram function is faster than repeated calls to SigWin using a range of window sizes.
w.c.deleeuw@uva.nl
SigWin-detector: a Grid-enabled workflow for discovering enriched windows of genomic features related to DNA sequences Marcia A Inda, Marinus F van Batenburg, Marco Roos, Adam SZ Belloum, Dmitry Vasunin, Adianto Wibisono, Antoine HC van Kampen, and Timo M Breit BMC Research Notes 2008; 1:63
Ridgeogram
lseq <- c(rnorm(200),rnorm(100,-1),rnorm(50),rnorm(50,4.0),rnorm(100)) plot((lseq-min(lseq))/(max(lseq)-min(lseq)),t="l") lines(18:(length(lseq)-17),SigWin(lseq,35),t="l",col="green") lines(18:(length(lseq)-17),SigWin(lseq,35,antiRidge=TRUE),t="l",col="red")