[BioC] cellHTS2: help with ConfFile (convertWellCoordinates) function

Wolfgang Huber huber at ebi.ac.uk
Wed Feb 27 23:57:46 CET 2008

Dear Dara,

thank you for sending the offending data files and the script you used. 
The solution is very simple: your file "Screenlog.txt" was empty (i.e. 
it only contained the header and no data rows), and the 'configure' 
function stumbled over that.

A work-around for you is to simply omit the 'logFile' argument from your 
call to 'configure' (as it is empty anyway). When starting R in the 
directory where the files are, the following lines worked for me:

x = readPlateList ('Platelist.txt', name='testcellHTSfiles')
x = configure(x, descripFile='Description.txt',

Also, we will fix the package so that it deals more gracefully with this 
'trivial' special case of zero-row Screenlog files.

Btw, in the code you had sent me, you had used 
dataPath=system.file(experimentName, package='cellHTS2'). You've 
probably seen this idiom in the vignette, where we use it to say where 
the "KcViab" example data are. While it is not wrong, it is unusual (and 
perhaps unnecessary) to copy your own, actual data into the directory 
where the R package is installed. Just put the data somewhere that makes 
sense to you, and set 'dataPath' to that path.

 > sessionInfo()
R version 2.7.0 Under development (unstable) (2008-02-27 r44632)


attached base packages:
  [1] splines   grid      tools     stats     graphics  grDevices utils
  [8] datasets  methods   base

other attached packages:
  [1] cellHTS2_2.3.19      vsn_3.4.8            limma_2.13.4
  [4] affy_1.17.5          preprocessCore_1.1.5 affyio_1.7.13
  [7] splots_1.5.1         genefilter_1.17.11   survival_2.34
[10] prada_1.15.0         rrcov_0.4-05         robustbase_0.2-8
[13] RColorBrewer_1.0-2   Biobase_1.17.14      fortunes_1.3-4

loaded via a namespace (and not attached):
[1] annotate_1.17.11     AnnotationDbi_1.1.22 DBI_0.2-4
[4] geneplotter_1.17.7   KernSmooth_2.22-22   lattice_0.17-6
[7] MASS_7.2-41          RSQLite_0.6-7        stats4_2.7.0

Best wishes

Wolfgang Huber  EBI/EMBL  Cambridge UK  http://www.ebi.ac.uk/huber

Wolfgang Huber scripsit:
> Dear Dara,
> the attachment didn't get through the mailing list, can you send it to 
> me privately (ideally, send a short example script that allows me to 
> reproduce your error, perhaps with some reduced or toy version of your 
> data) and I'll report the solution at this place.
> Generating valid (for read.table) tab-delimited tables with Excel is 
> remarkably difficult - does anyone on this list have a better suggestion 
> of how to let users create these files (a) with a tool they like and 
> know and (b) in a way that more robustly reads into R?
> 	Thanks and best wishes
> 		 Wolfgang
> ------------------------------------------------------------------
> Wolfgang Huber  EBI/EMBL  Cambridge UK  http://www.ebi.ac.uk/huber
> dunicadj at tcd.ie scripsit:
>> Hi all ,
>> I have a problem when configuring my files in CellHTS2. Primarily, I get an
>> error concerning my ConfFile which says the following-
>> Error in convertWellCoordinates(slog$Well, pdim(object)) :
>>   'x' must be either a character vector with alphanumeric well IDs (e.g. 'B03'
>> or c('B', '03'))
>>  or a vector of integers with position IDs within a plate (e.g. 27).
>> My Plateconf.txt file looks like this, which to me seems the right format.
>> However, regardless of whether I specify each plate and well as below, or use
>> the compressed format shown in the cellHTS2 vignette I still get the same
>> error.
>> Wells: 96
>> Plates: 18
>> Plate	Well	Content
>> 1	A01	Neg
>> 1	A02	Neg
>> 1	A03	Sample
>> 1	A04	Sample
>> 1	A05	Sample
>> 1	A06	Sample
>> My only naïve idea is that while generating the tab delinated txt file using
>> exell that the final well descriptions are not recognised as alphanumeric by
>> cellHTS2. I have tried generating the file in notepad but with no success.
>> Is there some other way I should be generating the plateConf file? I have even
>> adapted the ConfFile from the KcViab example folder and stil get the same
>> error.
>> I have attached a copy of my ConfFile if necessary
>> Any help would be greatfully appreciated.
>> Dara Dunican
>> Cell Signalling Group,
>> Institute of Molecular Medicine,
>> Trinity College Dublin,
>> Ireland

More information about the Bioconductor mailing list