This screening experiment was conducted to study how changing
treatment conditions would affect the yield of monoclonal antibodies
produced in mice innoculated with hybridoma cells. The response is
proportional to the number of antibody molecules produced. The
experimental design is a fractional factorial,
ff0616, that has 6
experimental factors each at two levels. The design consists of 16
runs and is a resolution IV, 1/4 fraction of the full factorial.
See Chapter 2 of Haaland (1989) for a more complete description and for a
follow-up response surface design. There is also a help file for the
follow-up experiment; namely,
abrsm.df.
ARGUMENTS:
RadDos
an experimental factor giving the radiation dose in rads.
Prime1
an experimental factor specifying
the time in weeks between the initial injection of Pristane
oil and the innoculation with antibody producing cells.
VolPrs
an experimental factor giving the volume of Pristane oil injected.
CelNum
an experimental factor showing
the number of antibody producing cells used in the innoculation.
Growth
an experimental factor indicating
the growth state of the antibody producing cells, either saturated
or log stage.
Prime2
an experimental factor indicating
whether or not a second priming with Pristane oil was used
immediately prior to innoculation with the antibody producing
cells
TtrVol
the response variable, antibody titer adjusted for volume. The
measured value
is proportional to the number of monoclonal antibody molecules
produced. The response should be maximized.
REFERENCES:
Haaland, P. D. (1989).
Experimental Design in Biotechnology.
New York: Marcel Dekker, Chapter 2.
SOURCE:
Used by permission of Marcel Dekker, Inc.
SEE ALSO:
EXAMPLES:
# This design is already available in S-PLUS
# under the name abscrn.df. These are the commands
# that were used to create it:
abscrn.fnames <- list(RadDos=c(250,500),Prime1=c(1,3),
VolPrs=c(0.1,0.5),CelNum=c('10e6','10e7'),Growth=
c('Log','Sat'),Prime2=c('No','Yes'))
abscrn.design <- design.digest(rep(2,6),abscrn.fnames,
fraction=~RadDos:Prime1:VolPrs:Prime2 +
RadDos:CelNum:Growth:Prime2+
Prime1:VolPrs:CelNum:Growth)
abscrn.TtrVol <- c(70,150,34,32,138,56,123,225,50,2.7,
1.2,12,90,2.1,4,15)
abscrn.df <- cbind(abscrn.design,TtrVol=abscrn.TtrVol)
# Sample analysis
plot(abscrn.df)
abscrn.fac <- fac.aov(abscrn.df)
summary(abscrn.fac)
pareto(abscrn.fac,sig=.10)
abscrn2.df <- abscrn.df[1:8,-1]
abscrn2.fac <- fac.aov(abscrn2.df)
pareto(abscrn2.fac,sig=.1)
qqnorm(abscrn2.fac,sig=.1)