OutlierD: Outlier dectection using quantile regression on the M-A...

In OutlierD: Outlier detection using quantile regression on the M-A scatterplots of high-throughput data

Description

This detects outliers using quantile regression on the M-A scatterplots of high-throughput data.

Usage

     OutlierD(x1, x2, k=1.5, method="nonlin")

Arguments

x1	one n-by-1 vector for data (n= number of peptides, proteins, or genes
x2	the other n-by-1 vector for data (n= number of peptides, proteins, or genes
k	parameter in Q1-k*IQR and Q3+k*IQR, IQR=Q3-Q1, k=1.5 (default)
method	one of constant, linear, nonlinear, and nonparametric quantile regression

Value

x	data and results for outliers

Author(s)

HyungJun Cho

Examples

data(lcms)
x <- log2(lcms) #log2-tranformation, do normalization if necessary

fit1 <- OutlierD(x1=x[,1], x2=x[,2], method="constant")
fit2 <- OutlierD(x1=x[,1], x2=x[,2], method="linear")
fit3 <- OutlierD(x1=x[,1], x2=x[,2], method="nonlin")
fit4 <- OutlierD(x1=x[,1], x2=x[,2], method="nonpar")

fit3$x[1:10,]

plot(fit3$x$A, fit3$x$M, pch=".", xlab="A", ylab="M")
i <- sort.list(fit3$x$A)
lines(fit3$x$A[i], fit3$x$Q3[i], lty=2); lines(fit3$x$A[i], fit3$x$Q1[i], lty=2)
lines(fit3$x$A[i], fit3$x$LB[i]); lines(fit3$x$A[i], fit3$x$UB[i])
title("Nonlinear")

OutlierD documentation built on Nov. 8, 2020, 6:12 p.m.