Using marrayRaw, Is there a way to combine two columns of data from a GenePix file into the Weight for a spot? For instance I would like to read in the "Flag" and "F Pixels" columns from the GenePix file to filter out all flagged spots and spots below a threshold size. It doesn't appear this is possible without modifying the marray classes. Thanks. -- Mike

Hi Mike, The current solution would be to modify read.marrayRaw and read in the two columns you are interested in. Construct a filter functions and store the results in the slot "maW". Alternately, you can store any statistics in the class "marrayTwo". See help(marrayTwo) and then construct a filter functions and store the results back in maW(marrayRaw). Hope this helps. Cheers Jean On Tue, 11 Nov 2003, Mike Schaffer wrote: > Using marrayRaw, Is there a way to combine two columns of data from a > GenePix file into the Weight for a spot? > > For instance I would like to read in the "Flag" and "F Pixels" columns > from the GenePix file to filter out all flagged spots and spots below a > threshold size. It doesn't appear this is possible without modifying > the marray classes. > > Thanks. > > -- > Mike > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor >

At 01:44 AM 12/11/2003, Mike Schaffer wrote: >Using marrayRaw, Is there a way to combine two columns of data from a >GenePix file into the Weight for a spot? > >For instance I would like to read in the "Flag" and "F Pixels" columns >from the GenePix file to filter out all flagged spots and spots below a >threshold size. It doesn't appear this is possible without modifying the >marray classes. read.maimages() in the limma package provides a way to do this. Simply define your own custom function mywtfun() which takes a data frame argument and computes the weights as you want, then use RG <- read.maimages(files, source="genepix", wt.fun=mywtfun) Converting the RGList object back to an marrayRaw object is pretty easy, if that's what you want to do, but you'd have to do it yourself. Gordon >Thanks. > >-- >Mike