agridat (version 1.5)

shafii.rapeseed: Rapeseed yield multi-environment trial, 3 years

Description

Rapeseed yield multi-environment trial, 3 years

Usage

data(shafii.rapeseed)

source

Electronic version retrieved from http://www.uiweb.uidaho.edu/ag/statprog/ammi/yld.data. Used with permission of Benjamin Price.

Details

SAS codes for the analysis can be found at http://www.uiweb.uidaho.edu/ag/statprog/ammi/ The data are from the U.S. National Winter Rapeseed trials conducted in 1986, 1987, and 1988. Trial locations included Georgia (GGA, TGA), Idaho (ID), Kansas (KS), Mississippi (MS), Montana (MT), New York (NY), North Carolina (NC), Oregon (OR), South Carolina (SC), Tennessee (TN), Texas (TX), Virginia (VA), and Washington (WA).

References

Bahman Shafii and William J Price, 1998. Analysis of Genotype-by-Environment Interaction Using the Additive Main Effects and Multiplicative Interaction Model and Stability Estimates, JABES, 3, 335--345.

Examples

dat <- shafii.rapeseed

dat$gen <- with(dat, reorder(gen, yield, mean))
dat$loc <- with(dat, reorder(loc, yield, mean))
dat$yield <- dat$yield/1000

dat <- transform(dat, rep=factor(rep), year=as.factor(as.character(year)))
dat$locyr = paste(dat$loc, dat$year, sep="")

# The 'means' of reps
datm <- aggregate(yield~gen+year+loc+locyr, data=dat, FUN=mean)
datm <- datm[order(datm$gen),]
datm$gen <- as.character(datm$gen)
datm$gen <- factor(datm$gen,
                       levels=c("Bienvenu","Bridger","Cascade",
                         "Dwarf","Glacier","Jet"))
dat$locyr <- reorder(dat$locyr, dat$yield, mean)

# This picture tells most of the story
# Now change symbols
op <- tpg <- trellis.par.get()
tpg$superpose.symbol$pch <- c('7','8','9')
trellis.par.set(tpg)
dotplot(loc~yield|gen,group=year,data=dat,auto.key=TRUE,
        ylab="Location")
dotplot(loc~yield|gen,group=year,data=datm,auto.key=TRUE)

# AMMI biplot.  Remove gen and locyr effects.
m1.lm <- lm(yield ~ gen + locyr, data=datm)
datm$res <- resid(m1.lm)
# Convert to a matrix
require("reshape2")
dm <- melt(datm, measure.var='res', id.var=c('gen', 'locyr'))
dmat <- acast(dm, gen~locyr)
# AMMI biplot.  Figure 1 of Shafii (1998)
biplot(prcomp(dmat))
trellis.par.set(op) # Unset graphics changes