Remove Color Channels from TIFF Using Command Line

5 08 2012

The R function


is used a lot to export R graphics e.g. for publication purposes. The tiff includes all 8 color channels and this collides with formatation requirements of most jounals (1200dpi but not more than some MB in size).

tiff("Whiskerplot.tiff" , width = 8.8 , height = 8.8 , units = "cm" , pointsize = 8 , res = 1200)
produces the required 1200 dpi resolution for monochrome plots, but with all 8 color channels (although just one is necessary) this results in 49.4MB for a tiny 8.8×8.8 cm (ca 3.5×3.5 inch) plot. I used GIMP to remove the color channels with

Image > Mode > Indexed > 1-bit monochrome

but this required opening GiMP, the file, the menu and saving. And again, when the plot had to be modified.

Since this has to be done for each and every publication figure I looked up a command line solution and here it is:
convert Whiskerplot.tiff -flatten -monochrome Whiskerplot_monochrome.tiff
removes all unnecessary colorchannels from Whiskerplot.tiff and saves it as Whiskerplot_monochrome.tiff, reducing the size from 49.4MB to 2.1MB while the result is exactly the same.

Hattip: Paddy Landau on ubuntuforums

R function to transform continuous variable to categorical factor cut at n-tiles

6 11 2011

The cut() function can be used to transform a continuous variable into a categorical factor variable. The syntax is quite lengthy and if one wishes to cut at quartiles, quintiles or other n-tiles one has to include the quantile() function into the call.

This is not very newbee friendly and if included into a model-call nearly unreadable.

The function in the code box cutN() does the job.

 cutN <- function(X , n = 4){
         X ,
         include.lowest = TRUE ,
         breaks = quantile(
                           X , 
                           probs = (0:n)/n ,
                           na.rm = TRUE ))}

In order to cut the continuous variable Creatinine in the dataset Patients into deciles (n=10) the syntax is:
cutN( Patients$Creatinine , n = 10 )

No big deal, but maybe useful…

Add public key behind a firewall in Ubuntu Shell

7 09 2011

In short: Use
sudo apt-key adv --keyserver hkp:// --recv-key E084DAB9
instead of
sudo apt-key adv --keyserver --recv-key E084DAB9
This way you force port 80 which is usually clear.

I got the idea from the answer of Phil Bradley on the forum. He claimed that this would be fixed in Natty, but it isn’t although the configuration file he mentions has the port80 specification added by default, apt-key does not use it. The above snippet solves that.

For those Ubuntu users who have no idea what I am talking about:

Installing the newest R-version in Ubuntu requires to append the CRAN repository to you /etc/apt/sources.list. One might hit Alt+F2 and enter
gksu gedit /etc/apt/sources.list

With Xubuntu you would use mousepad instead of gedit. In any distro you can use
sudo nano /etc/apt/sources.list
in a terminal.

Usually I add the line
deb natty/
at the end of the file and update with
sudo apt-get update.

CRAN at University of Bergen is closest to me. You might want another one (check the site for mirrors).

apt-get update answers with a warning
GPG error: nat/ Release: The folowing signatures coldn't be verified because the public key is not abailable

That is not a problem. One can install R and packages anyway, but it is better to have the public key.

Behind a firewall (and many public and open hotspots block several ports) it is not possible to use

sudo apt-key adv --keyserver --recv-key E084DAB9

since the port through which the keyserver is contacted is blocked on most firewalls. You have to force port 80 by:
sudo apt-key adv --keyserver hkp:// --recv-key E084DAB9

After the key is added
sudo apt-get update
sudo apt-get install R-recommended emacs ess

proceeds without warning nor error.

Mathematical expressions in R plot

3 03 2011

Mathematical anotations to R plots can be formated LaTeX style with the expression function. The expression() can be included in

  • plot titles
    title( expression(...))
  • axis anotations
    plot( ... , xlab = expression(...))


  • the plot panel itself
    text( x , y , expression(...))

An example:

plot( 0 , 0 , type = "n" , xlab = expression( "Nothing" * (mu*mol/l)) )
text( 0 , 0 , expression(beta>=0.2))
title( expression( "Only an expression() demo " * theta^(2*pi)))

It is even possible to update the plot anotation from a variable (found on the R forum)

[R] Expression in plot text

Roger Koenker roger at
Wed Dec 6 20:22:05 CET 2000

expression(paste(hat(theta),'= ',that)))

We’ve been here before. To get an expression either use parse on your
pasted character string or substitute on an expression. There are worked
examples in the list achives. The neatest is

title(substitute(hat(theta) == that, list(that=that)))

(note it is == not =)

Violinplots with R

7 02 2011

Just for the record a nice alternative to boxplot() from the R introduction on MOAC@University of Warwick:

Violin plots are essentially pretty versions of box plots, where the width is set by the local density. For skewed distributions, you get things that look a bit like “violins”, hence the name.

The catch is that we need to install and load the package UsingR to get the violin plot function. […]
simple.violinplot(Age ~ Gender, data=Santa, col = "red")

Plot Function For Additive Cox Proportional Hazard Regression

30 10 2009

Note: This initial blog post is discontinued. If you want to have the latest development of the function have a look at the static page PlotHR

Usually a Cox-regression is achieved in R by
model <- coxph ( SuvivalObject ~ Covariate1 + Covariate2 + Factor1 + Factor2 , data = Dataset )

The covariates can be enclosed in other funtions:

  • factors should be enclosed by factor()
  • strata, which allow to adjust for a factor without getting an estimate, should be enclosed by strata()
  • non-log linear continuous terms can be enclosed by

In the latter case the model might look like
model <- coxph (SurvivalObject ~ pspline(Covariate1) + Covariate2 + factor(Factor1) + strata(Factor2) , data = Dataset )

The functional form of the covariates (including the factors) can now be plotted with

Though the termplot() function fails with plotting just one covariate and leaves no cusomization.

The function plotHR() plots the functional form of the desired term: plotHR(model)
plots the first term in the model by default but other terms can be accessed by calling their number (e.g. the second one):
plotHR(model , terms = 2)

In order to use the function you have to “source” it into R. It is the same procedure as calling a package, but using “source” instead of “library”.

Paste the function syntax into a textfile and safe it (as plot.HR.R) on your harddisk, remember the path and include
before using the function.

Download plotHR()

Note: I have rewritten the function several times since I wrote the initial post … using version numbers now …

  • V0.6 – removed the y.log option, since the scale should be logaritmic anyway. Later I will also rewrite the x.log option, since the feature is already incorporated in the plot.default() function. I also removed the dottet line at HR=1 level, since some complained about it overstating the importance of the log(HR) intercept. I included it, since it gives a hint about the significance of the smooth term, in case the confidence intervalls cross over the line… Those who miss it can add manually lines( h = 0 , type = 2 )

    I rewrote the “rugs” option. Try rugs = "density" It is still “beta”ish, but some like it.

  • V0.5 – bug fix for the y-scale and slight adjustment of the default plotting colors (paler CI shade and stronger term-line)
  • V0.4 – the y-scale should be logarithmic; a HR of 0.5 (50% reduced Hazard) should show the same distance from HR = 1 as a doubled Hazard (HR = 2); this is now default. The linear scale I used initially is biased in this concern (Hat-tip: Arve Ulvik, Eva Pedersen and Roy Nilsen). The option y.log allows both ways (linear and log-scale); the axis labels denote Hazard Ratio instead of log(HR).

plotHR( model , terms = 1 , se = TRUE , rug = "ticks" , x.log = FALSE , xlab = "" , ylab = "Hazard Ratio" , main = NULL , xlim = NULL , ylim = NULL, col.term = "#08519C", lwd.term = 3, = "#DEEBF7", cex = 1 , bty = "n" , axes = TRUE )

  • model – a coxph model
  • terms – integer; the number of the term to plot
  • se – logical TRUE/FALSE; plotting the CI
  • rug – “ticks” or “density”; rug plot or density plot at x-axis. Any other value for “rug” will omit the rugplot.
  • x.log – logical TRUE/FALSE; log-transformed exposure variable
  • xlab – character; x-axis label
  • ylab – character; y-axis label
  • main – character; main plot title
  • xlim – 2×1 column vector; x-range of plot
  • ylim – 2×1 column vector; y-range of plot
  • col.term – color of HR-curve
  • lwd.term – line width of HR-curve
  • – color of CI (if plotted)
  • cex – numeric; size factor of labels
  • bty – specifies the boxtype around the plot. See ?plot.default

Installation of Package ‘rggobi’ requires GGobi

30 08 2009

Installation of package ‘rggobi’ had non-zero exit status on my way to solve all the rattle dependencies.

rggobi is an interface from R to GGobi – an open source visualization program for exploring high-dimensional data. It provides highly dynamic and interactive graphics such as tours, as well as familiar graphics such as the scatterplot, barchart and parallel coordinates plots. Plots are interactive and linked with brushing and identification (from the package description).

When trying to install rggobi R complained accordingly error: Package requirements (ggobi >= 2.1.6) were not met: No package 'ggobi' found

and the fix is not hard to figure out:
sudo aptitude install ggobi

rggobi also depends on RGtk2 which in turn requires libglade. That was another post.