[BioC] limma - reading data from different Print files in cDNA

Jenny Drnevich drnevich at uiuc.edu
Fri Jan 12 18:31:02 CET 2007

Hi Dipen,

If I am understanding you correctly, you have two different sets of gpr 
files. Both sets have the same gene/spot IDs, but they were printed in a 
different order, correct?  I'm not exactly sure how read.maimages reads in 
the $genes information (i.e., block, row, column, id, etc.), but it might 
just read it from the first gpr file. If so, the intensities for all the 
gpr files of this type should be correct, but the other .gpr files will be 
incorrect. If not, what do you mean when you say that it is "collating the 
gene-intensity information" ?

You could manually go into the gpr files of one type and sort the spot IDs 
so they are in the same order as the other gpr type, but I wouldn't 
recommend this approach. Because the two gpr files have different spatial 
arrangements of the genes, any normalization that takes into account 
spatial location (e.g., print-tip loess within-array normalization) would 
be done incorrectly for the re-ordered gpr files. The second, better 
approach would be to read in each set of gpr files separately, do any 
spatially-dependent normalization separately (which are done per array, so 
doing them in separate groups won't matter), then sort them so they both 
have the same spot order and merge them together using merge(). You might 
have to replace the block, row, column info of one set with the other set 
for merge() to work properly, but as long as it's AFTER doing any 
spatially-based corrections, it will be fine. This second approach will 
also work well if there are differences in the total number of spots or 
numbers of blanks, buffers, etc. You'll have to remove all the blank and 
buffer spots from each set before merging.


At 09:43 AM 1/11/2007, Dipen Sangurdekar wrote:
>When I attempt to read in different gpr files (originating from slightly
>different print files, but with same gene IDs) for the same experimental
>setup, limma makes errors in compiling the RG file. The wrong intensity
>values are associated with a particular gene across the arrays.
>Is there a quick way to get around the problem, that avoids collating
>gene-intensity information from each gpr?
>         [[alternative HTML version deleted]]
>Bioconductor mailing list
>Bioconductor at stat.math.ethz.ch
>Search the archives: 

Jenny Drnevich, Ph.D.

Functional Genomics Bioinformatics Specialist
W.M. Keck Center for Comparative and Functional Genomics
Roy J. Carver Biotechnology Center
University of Illinois, Urbana-Champaign

330 ERML
1201 W. Gregory Dr.
Urbana, IL 61801

ph: 217-244-7355
fax: 217-265-5066
e-mail: drnevich at uiuc.edu

More information about the Bioconductor mailing list