gamair (version 1.0-2)

prostate: Prostate cancer screening data

Description

Protein mass spectographs for patients with normal, benign enlargement and cancer of the prostate gland.

Usage

data(prostate)

Arguments

Format

A three item list

type

1 for normal, 2 for benign enlargement and 3 for cancerous.

intensity

A matrix with rows corresponding to measurements in type. Each row is a normalized spectral intensity measurement for the protein mass given in MZ

MZ

Matrix corresponding to intensity giving the protein masses in Daltons.Actually all rows are identical.

Details

See the source article for fuller details. The intensity data here have been smoothed so that each measurement is an average of 40 adjacent measurements from the raw spectrum. The intensity data have also been rounded to 3 significant figures. This pre-processing was done to reduce the dataset size to something reasonable for distribution.

References

Adam, B-L. Y. Qu, J.W. Davis et al. (2002) Serum Protein Fingerprinting Coupled with a Pattern-matching Algorithm Distinguishes Prostate Cancer from Benign Prostate Hyperplasia and Healthy Men. Cancer Research 62:3609-3614

Examples

Run this code
# NOT RUN {
require(gamair);require(mgcv)
data(prostate)
## plot some spectra...
par(mfrow=c(2,3),mar=c(5,5,3,1))
ind <- c(1,163,319)
lab <- list("Healthy","Enlarged","Cancer")
for (i in 1:3) {
  plot(prostate$MZ[ind[i],],prostate$intensity[ind[i],],type="l",ylim=c(0,60),
  xlab="Daltons",ylab="Intensity",main=lab[[i]],cex.axis=1.4,cex.lab=1.6)
  lines(prostate$MZ[ind[i],],prostate$intensity[ind[i]+2,]+5,col=2)
  lines(prostate$MZ[ind[i],],prostate$intensity[ind[i]+4,]+10,col=4)
}
## treat as ordered cat control, bph, cancer
b <- gam(type ~ s(MZ,by=intensity,k=100),family=ocat(R=3),
         data=prostate,method="ML")
## results...
pb <- predict(b,type="response")
plot(b,rug=FALSE,scheme=1,xlab="Daltons",ylab="f(D)",
cex.lab=1.6,cex.axis=1.4,main="a")
plot(factor(prostate$type),pb[,3],cex.lab=1.6,cex.axis=1.4,main="b")
qq.gam(b,rep=100,lev=.95,cex.lab=1.6,cex.axis=1.4,main="c")
# }

Run the code above in your browser using DataCamp Workspace