computeFDRwithID: Compute FDR given a score

In prozor: Minimal Protein Set Explaining Peptide Spectrum Matches

Description

For more details and references see package vignette vignette("TargetDecoyFDR_Example", package = "prozor")

Usage

computeFDRwithID(score, ID, decoy = "REV_", larger_better = TRUE)

Arguments

score	a vector with scores
ID	- list with protein id's
decoy	decoy pattern, default "REV_"
larger_better	if larger score better than small (default TRUE), If small score better set FALSE

Value

list with ID, decoy_hit (indicates if decoy), score the search engine score, FDR1 false discovery rate estimated using the method of Elias and Gygi; FDR2 - estimated using the method of Kell.

Examples

data(fdrSample)
# call constructor
#nrow(fdrSample)
#fdrSample <- dplyr::slice_sample(fdrSample, n = 40000)

#usethis::use_data(fdrSample, overwrite = TRUE)
fdr1 <- computeFDRwithID(fdrSample$score, fdrSample$proteinID, larger_better = FALSE)
names(fdr1)
plot(fdr1$score, fdr1$FPR,type="l",xlim=c(0,0.001), ylim=c(0,0.0002))
lines(fdr1$score, fdr1$qValue_FPR, col=2)
lines(fdr1$score, fdr1$SimpleFDR,type="l",col=4)
lines(fdr1$score, fdr1$qValue_SimpleFDR, col=5)


fdr1 <- computeFDRwithID(fdrSample$score2, fdrSample$proteinID, larger_better = TRUE)
names(fdr1)
plot(fdr1$score, fdr1$FPR,type="l", xlim=c(2.5,5),ylim=c(0,0.001))
lines(fdr1$score, fdr1$qValue_FPR, col=2)
lines(fdr1$score, fdr1$SimpleFDR,type="l",col=4)
lines(fdr1$score, fdr1$qValue_SimpleFDR, col=5)

prozor documentation built on Dec. 11, 2021, 9:51 a.m.