Hi Luigi
If you str(stripplot(.)) ie your example you will see that the x is
categorical and it is harder to put lines onto it than xyplot so that is
what I referred to
Try this - I have added a column and put in segments - I converted it to
xyplot because of above
I have set the x axis limits so if you change them the segments may need
changing
abline works OK as it does not have reference to the x axis
I deleted your pch argument and modified the par.settings to suit
I have not checked the values but may be necessary to put in extra lines of
code for the 2 "groups"
you may now be able to work with groups = TB and modify to suit if you need
to work with groups
see ?panel.superpose
?panel.groups
dat <-
structure(list(patient = c("A264", "A264", "A264", "A264", "A264",
"A264", "A264", "H041", "H041", "H041", "H041", "H041", "H041",
"H041", "A263", "A263", "A263", "A263", "A263", "A263", "A263",
"N241", "N241", "N241", "N241", "N241", "N241", "N241"), TB = c("A",
"A", "A", "A", "A", "A", "A", "L", "L", "L", "L", "L", "L", "L",
"A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A",
"A"), gene = c("IP10", "IP10", "IP10", "IP10", "IP10", "IP10",
"IP10", "IP10", "IP10", "IP10", "IP10", "IP10", "IP10", "IP10",
"IP10", "IP10", "IP10", "IP10", "IP10", "IP10", "IP10", "IP10",
"IP10", "IP10", "IP10", "IP10", "IP10", "IP10"), stimulus = structure(c(1L,
2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L,
4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L), .Label = c("ESAT6",
"CFP10", "Rv3615c", "Rv2654", "Rv3879", "Rv3873", "PHA"), class = "factor"),
ratio = c(250.9, 399.1, 959.4, 1.994, 589.2, 5.366, 48.27,
10.32, NA, 3.33, NA, 1, NA, 1, 509.5, 991, 689.6, 1285, 683.7,
468.7, 47.08, 8.534, 2.733, NA, 11.18, 10.41, 0.7532, NA)), .Names =
c("patient",
"TB", "gene", "stimulus", "ratio"), row.names = c(NA, -28L), class =
"data.frame")
dat$x <- ifelse(dat$TB == "A",1,2)
datavA <- aggregate(ratio ~ TB+stimulus, dat, mean, na.rm = T)
datmeA <- aggregate(ratio ~ TB+stimulus, dat, median, na.rm = T)
xyplot(ratio ~ x|stimulus, dat,
hor=F,
as.table = TRUE,
layout = c(7,1),
ylab=expression(bold("Ratio GOI/RG")),
xlab=expression(bold("Antigen stimulation")),
main="IP10 expression",
scales = list(x = list(alternating = FALSE,
at = c(1,2),
labels = c("A","L"))),
xlim = c(0,3),
col=c("red","black"),
par.settings = list(strip.background=list(col="white"),
superpose.symbol = list(col =
c("red","black"),
pch = c(19,19))),
par.strip.text=list(font=2),
panel = function(x, y, ...){
pnl = panel.number()
# mean
panel.segments(x0 = x-0.25,
y0 = datavA[datavA[,2]==
levels(datavA[,2])[pnl],"ratio"],
x1= x+0.25, y1 = datavA[datavA[,2]==
levels(datavA[,2])[pnl],"ratio"], col = c("red","black"))
# median
panel.segments(x0 = x-0.25,
y0 = datmeA[datmeA[,2]==
levels(datmeA[,2])[pnl],"ratio"],
x1= x+0.25, y1 = datmeA[datmeA[,2]==
levels(datmeA[,2])[pnl],"ratio"], lwd = 2, col = c("magenta","grey"))
panel.stripplot(x,y, ...)
}
)
Duncan
From: Luigi Marongiu [mailto:marongiu.lu...@gmail.com]
Sent: Sunday, 26 January 2014 03:58
To: Duncan Mackay
Subject: Re: [R] add median/mean line to stri-plot in lattice
Dear Duncan,
sorry to pester you again with this issue. The figure "strip_chart" was
obtained using the basic R setting and that was perfectly satisfactory for
its purpose; now I need to expand such approach to the "result database",
which is more complex than the one that generated "strip_chart".
In the code you provided me the lines are correctly drawn, only thing I need
to make them shorter as in the "strip_chart" example, a task that I obtained
with the code: segments(x-0.1, y, x+0.1, lwd=2) where y was the vector
containing the median values of the data.
What would be the equivalent for Lattice? I reckon it should be done with
replacing panel.abline with something else.
Many thanks again. And thank you also for reference, I will study it but it
might take sometimes.
Best regards
Luigi
stripplot(ratio ~ TB|stimulus, dat,
hor=F,
as.table = TRUE,
layout = c(7,1),
scales = list(relation = "same"),
ylab=expression(bold("Ratio GOI/RG")),
xlab=expression(bold("Antigen stimulation")),
main="IP10 expression",
pch=19,
col=c("red","black"),
par.settings = list(strip.background=list(
col="white"),
superpose.symbol = list(col =
c("red","black"),
pch = c(1,20))) ,
par.strip.text=list(font=2),
panel = function(x, y, ...){
pnl = panel.number()
# mean
panel.abline(h = datavA[datavA[,2]==
levels(datavA[,2])[pnl],"ratio"], col = c("red","black"))
# median
panel.abline(h = datmeA[datmeA[,2]==
levels(datmeA[,2])[pnl],"ratio"], col = c("red","black"), lty=3)
panel.stripplot(x,y, ...)
}
)
On Sat, Jan 25, 2014 at 11:45 AM, Duncan Mackay <dulca...@bigpond.com>
wrote:
Hi Luigi
If you create a numeric variable of 1:5 to correspond as the x's - Normal,
Borderline, .
You could then plot the points as an xyplot -xyplot gives you more
flexibility to do things.
then add the lines as segments in a panel function
Use scales for the labels as
scales = list(x = list(at = 1:5, labels = c("Normal","Borderline", .))),
Regards
Duncan
From: Luigi Marongiu [mailto:marongiu.lu...@gmail.com]
Sent: Saturday, 25 January 2014 00:45
To: Duncan Mackay
Subject: Re: [R] add median/mean line to stri-plot in lattice
Dear Duncan,
sorry for the data, it was a bit too bulky to simply put it here as a table,
although you recreated it very good.
What i would like to do is something similar to the figure attached, which
was done with the normal strip chart. For each group there is the line
indicating the median value.
In the query I had the data is more complex and therefore "lattice" was a
better choice, but the syntax is very difficult, in fact you had to
implement a function whithin the function stripplot.
Essentially there should be a short line overlapping each data set
representing the median (or average) values. Probably it will be a matter of
changing panel.abline, but I don't know how.
Thank you very much for your help
Luigi
On Thu, Jan 23, 2014 at 11:45 PM, Duncan Mackay <dulca...@bigpond.com>
wrote:
Hi Luigi
Not sure what you want exactly.
I have cleaned up the formula - do that at the data.frame first.
see http://lmdvr.r-forge.r-project.org/figures/figures.html for more
examples on lattice
Please dput your data - makes things easier and we have what you are working
with
There may be better ways of doing this but try
dat <-
structure(list(patient = c("A264", "A264", "A264", "A264", "A264",
"A264", "A264", "H041", "H041", "H041", "H041", "H041", "H041",
"H041", "A263", "A263", "A263", "A263", "A263", "A263", "A263",
"N241", "N241", "N241", "N241", "N241", "N241", "N241"), TB = c("A",
"A", "A", "A", "A", "A", "A", "L", "L", "L", "L", "L", "L", "L",
"A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A",
"A"), gene = c("IP10", "IP10", "IP10", "IP10", "IP10", "IP10",
"IP10", "IP10", "IP10", "IP10", "IP10", "IP10", "IP10", "IP10",
"IP10", "IP10", "IP10", "IP10", "IP10", "IP10", "IP10", "IP10",
"IP10", "IP10", "IP10", "IP10", "IP10", "IP10"), stimulus = structure(c(1L,
2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L,
4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L), .Label = c("ESAT6",
"CFP10", "Rv3615c", "Rv2654", "Rv3879", "Rv3873", "PHA"), class = "factor"),
ratio = c(250.9, 399.1, 959.4, 1.994, 589.2, 5.366, 48.27,
10.32, NA, 3.33, NA, 1, NA, 1, 509.5, 991, 689.6, 1285, 683.7,
468.7, 47.08, 8.534, 2.733, NA, 11.18, 10.41, 0.7532, NA)), .Names =
c("patient",
"TB", "gene", "stimulus", "ratio"), row.names = c(NA, -28L), class =
"data.frame")
#dat$stimulus = factor(dat$stimulus, levels =
c("ESAT6","CFP10","Rv3615c", "Rv2654", "Rv3879","Rv3873","PHA"))
datav <- aggregate(ratio ~ TB, dat, mean, na.rm = T)
datme <- aggregate(ratio ~ TB, dat, median, na.rm = T)
datavA <- aggregate(ratio ~ TB+stimulus, dat, mean, na.rm = T)
datmeA <- aggregate(ratio ~ TB+stimulus, dat, median, na.rm = T)
stripplot(ratio ~ TB|stimulus, dat,
hor=F,
as.table = TRUE,
layout = c(7,1),
scales = list(relation = "same"),
ylab=expression(bold("Ratio GOI/RG")),
xlab=expression(bold("Antigen stimulation")),
main="IP10 expression",
pch=19,
col=c("red","black"),
par.settings = list(strip.background=list(col="white"),
superpose.symbol = list(col =
c("red","black"),
pch = c(1,20))) ,
par.strip.text=list(font=2),
panel = function(x, y, ...){
pnl = panel.number()
# mean
panel.abline(h = datavA[datavA[,2]==
levels(datavA[,2])[pnl],"ratio"], col = c("red","black"))
# median
panel.abline(h = datmeA[datmeA[,2]==
levels(datmeA[,2])[pnl],"ratio"], col = c("red","black"), lty=3)
panel.stripplot(x,y, ...)
}
)
stripplot(ratio ~ TB|stimulus, dat,
hor=F,
as.table = TRUE,
layout = c(7,1),
scales = list(relation = "same"),
ylab=expression(bold("Ratio GOI/RG")),
xlab=expression(bold("Antigen stimulation")),
main="IP10 expression",
pch=19,
col=c("red","black"),
par.settings = list(strip.background=list(col="white"),
superpose.symbol = list(col =
c("red","black"),
pch = c(1,20))) ,
par.strip.text=list(font=2),
panel = function(x,y,...){
panel.abline(h = datav[,2], col = c("red",
"black"),lty=1)
panel.abline(h = datme[,2], col = c("red",
"black"),lty=3)
panel.stripplot(x,y, ...)
}
)
Regards
Duncan
Duncan Mackay
Department of Agronomy and Soil Science
University of New England
Armidale NSW 2351
Email: home: mac...@northnet.com.au
-----Original Message-----
From: r-help-boun...@r-project.org [mailto:r-help-boun...@r-project.org] On
Behalf Of Luigi Marongiu
Sent: Friday, 24 January 2014 06:55
To: r-help@r-project.org
Subject: [R] add median/mean line to stri-plot in lattice
dear all,
I set a series of variable subdivided in classes (there are 7 in here,
defined by the clause 'levels'), and I am showing the results in strip
chart using the package 'lattice'. Whitin each class the samples are
further subdivided in "active" (A) and "latent" (L). The function I have
written is:
stripplot(my.data.IP10$ratio ~
factor(my.data.IP10$TB)|factor(my.data.IP10$stimulus,
levels = c("ESAT6","CFP10","Rv3615c", "Rv2654", "Rv3879",
"Rv3873","PHA")),
my.data, hor=F, layout = c(7,1), scales = list(relation = "same"),
ylab=expression(bold("Ratio GOI/RG")), xlab=expression(bold("Antigen
stimulation")), main="IP10 expression", pch=19, col="black",
par.settings = list(strip.background=list(col="white")),
par.strip.text=list(font=2))
which defines the figure attached "IP10".
Providing the median or average values (that will depend on the
distribution of the final data), is there a way to add a line overlapping
each of these classes and subclasses? I reckon it will be a vector of 14
elements, which can be called "Mvect".
In addition is there a way to differently colour the subclasses? lest say
black for active and white for latent? I tried with 'group' but with not
success.
Thank you for any help, I am not quite familiar with the Latex codes. I am
also attaching the dataframe containing the data for further reference.
Best regards
Luigi
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
