[BioC] ArrayVision and limma

Gordon Smyth smyth at wehi.edu.au
Tue Aug 26 22:35:45 MEST 2003

Sorry, I have to withdraw my statement that 'read.maimages' can read 
ArrayVision files. Looking at the attachment from Isaac, it seems that 
ArrayVision writes column names over two rows of the file, and in addition 
these column names include non-ascii characters.  'read.maimages' can't 
handle a non-standard feature like this without special programming.

However, since Jean has apparently just written a function to read 
ArrayVision data into marray, the obvious solution is to use her function.


At 03:50 PM 26/08/2003, Gordon Smyth:
>At 12:12 PM 26/08/2003, Isaac Mehl wrote:
>>I just received some wonderful help from Jean on reading ArrayVision
>>output files into marrayInput. Now i would like to try and analyze with
>>limma.  if i understand correctly there is no way to do this unless
>>source = "ArrayVision" which doesn't exist.  Read:
>>which leads me to believe that the only way to accomplish this is for
>>Gordon to write a separate function.  Is this the only way to read in
>>data?  Won't this be a pain with many, many different software packages
>>making spot calls??  Isn't it possible to simply specify which column
>>contains the needed information (RF, Gf, Rb, etc) in a matrix??
>I don't see the problem. If you've read in the data using marrayInput and 
>then normalized using marrayNorm, you can use any of the linear model 
>functions in limma to analyse your data, as explained in the section 
>called "Using limma with the marray Packages" in the LIMMA User's Guide. 
>The easiest function to use would be 'lmFit', in the development version 
>of limma, because it accepts 'marrayNorm' objects directly.
>If for some reason you want to read in your data using limma instead of 
>marrayNorm, you could do so by using the 'columns' argument of 
>'read.maimages' to specify which columns of your data files correspond to 
>Rf, Gf, Rb and Gb. This option has always been available in limma, 
>although the development version of limma will do a better job of handling 
>column names containing spaces, like the names in your file. As the help 
>document for 'read.maimages' says,
>    "At present, the image analysis programs SPOT, GenePix, QuantArray
>    and Imagene are supported. Other image analysis programs can
>    potentially be accommodated though a user-specified 'columns' argument."
>The reason that Imagene data required special intervention from me was (i) 
>because Imagene is almost unique in that it writes the red and green 
>channel information to separate files and (ii) because neither limma nor 
>marrayNorm supported imagene explicitly.
>>I am attaching an example of a .csv output from ArrayVision if anyone
>>would like a crack at it.
>>-isaac mehl
>>gene expression lab (gele)
>>salk institute
>>10010 n. torrey pines rd.
>>la jolla ca. 92037

More information about the Bioconductor mailing list