Progress update (summarized from my forum for such matters at http://www.theschedule.net/forum/gforum.cgi?forum=20&do=forum_view): Briefly, I created a four-node cluster out of Pentium-III boxes and Debian Linux/openMosix. I saw no significant performance boost of ReadAffy or expresso using the set of 165 .CEL files from Harvard. None of the processes migrated, as they say in the world of high- performance computing. R.bin runs in one process, and everything it does seems to stay in that process. No real opportunity for parallelization here, at least not on openMosix. I'd like to analyze these chips in a reasonable amount of time, without paying Dell $45,000 for 4-Xeon SMP server. I worry what we'll do with 1,000 .CEL files. The analytical techniques work well, but pretty slow even if your amp "goes to 11." Any thoughts? Michael Benjamin, MD Emory University

Hi, I'm able to read the Agilent flat file using the read.table function but what if I have a large amount of arrays to input. This function seems to only input one table at a time. Are there any ways to read in multiple files simultaneously (as in the read.maimages function)? Can you please list the commands/functions I would need to perform the reading in normalized signals task? Then I can go and look through help in R for them? How would I go about defining the weight matrix for each array since the source for the Agilent analysis program is not included in the read.maimages function? I have a column of filter values in the Agilent txt file which was read in through read.maimages (manually through the column argument). How can I define this filter column? Thanks for you time, Anna -----Original Message----- From: Gordon Smyth [mailto:smyth@wehi.edu.au] Sent: Wednesday, December 10, 2003 4:45 PM To: Anna Cao Cc: bioconductor@stat.math.ethz.ch Subject: Re: [BioC] limma package At 06:15 AM 11/12/2003, Anna Cao wrote: >Hi, > >I'm trying to normalize between arrays using the limmma package so I >can run SAM. I want to know whether we can read log2 ratio data into >R? I'm using Agilent's extraction program to extract cDNA arrays >data and the image analysis program from Agilent automatically >normalize within array, giving the normalized intensity and log >ratio. Is there anyway I can by-pass reading Mean/Median foreground >and background intensity and directly feed the processed (background >subtracted and normalized) intensity into R? This is easy in R. Use the Agilent flat file, use read.table(), then collect the normalized columns into a matrix in R. The matrix can then but used as input for example to the lmFit() command in limma. You'll have to do the input yourself though, there aren't any canned limma commands to do it. If you need help with this, you might want to ask for help on the R-help mailing list. >Another question: Can I also input a column that filters spots which >were unsatisfactory according to the image analysis program? And how >can I can remove these bad spots from further calculation in R using >the limma package or any other functions in R? Using limma, you set the appropriate elements of the 'weights' matrix to zero for those spots. This applies to both normalization and linear modelling functions in limma. Normally the weights matrix is set automatically using the wt.fun argument of read.maimages, but you will need to create it yourself if you're reading in the data manually as it were. Gordon >Thanks, > >Anna