In lshen1/drexplorer2: Dose-response Explorer2

knitr::opts_chunk$set(
  message = FALSE,
  digits = 3,
  collapse = TRUE,
  comment = "#>"
  )
options(digits = 3)

This package implements the method by Newman, D. The test statistic q=w/s where w is the range of the data and s is sample standard deviation estimated from controls. The null distribution for q has been derived and 1% and 5% quantiles have been given in the paper.

We implement this procedure. In particular, NewmanTest() returns a logic vector specifying whether the observations are outliers.

Usually, drug-response data contains multiple doses. Therefore, we write a wrapper drOutlier() that compute the result for all doses one dose at a time.

We use the ryegrass data from drc package for illustration purpose. The ryegrass data was originally published by (Streibig et al. 2002) which contains 24 concentrations of ferulic acid (a root growth inhibitor) and corresponding root length.

First, we load the drexplorer package and attach the \Rpackage{ryegrass} data.

library(drexplorer2)

data(ryegrass)
dose <- ryegrass[, 2]
response <- ryegrass[, 1]

At dose=3.75 and significance level 0.05, we find there is one outlier identified:

## potential outlier at dose 3.75
NewmanTest(ref = response[dose == 0], obs = response[dose == 3.75], alpha  =0.05)

We also examine all dose levels and find no further outliers:

drOutlier(drMat = ryegrass[, c(2, 1)], alpha = 0.05)

lshen1/drexplorer2 documentation built on June 2, 2020, 9:27 p.m.