I've managed to use genefilter to generate an array of FALSE and TRUE's for the affy bovine array I'm working with: -- start code -- # expSet.RNA.RMAqndata is an ExpressionSet object with log2 expression values f1<-anyNA f2<-kOverA(16, log2(100)) # filter out genes with low/no expression over all arrays ff<-filterfun(f1,f2) wh<-genefilter(expSet.RNA.RMAqndata, ff) sum(wh) # number of genes passing the filter -- end code -- I was wondering if someone could shed light on how to use the wh array downstream with things such as limma etc. Thanks Nathan ------------------------------------------------------------- Dr. Nathan S. Watson-Haigh (publish under Haigh, N.S.) OCE Post Doctoral Fellow CSIRO Livestock Industries J M Rendel Laboratory Rockhampton QLD 4701 Tel: +61 (0)7 4923 8121 Australia Fax: +61 (0)7 4923 8222

Hi Nathan, Nathan.Watson-Haigh at csiro.au wrote: > I've managed to use genefilter to generate an array of FALSE and TRUE's > for the affy bovine array I'm working with: > > -- start code -- > # expSet.RNA.RMAqndata is an ExpressionSet object with log2 expression > values > > f1<-anyNA > f2<-kOverA(16, log2(100)) # filter out genes with low/no expression > over all arrays > ff<-filterfun(f1,f2) > wh<-genefilter(expSet.RNA.RMAqndata, ff) > sum(wh) # number of genes passing the filter > -- end code -- > > I was wondering if someone could shed light on how to use the wh array > downstream with things such as limma etc. > What you have isn't an array - it's a vector of TRUE/FALSE that you can use to subset your ExpressionSet to remove the reporters that don't pass the filter. expSet.RNA.RMAqndata.filtered <- expSet.RNA.RMAqndata[wh,] And then you use the filtered data with limma, etc. Best, Jim > > Thanks > Nathan > > > ------------------------------------------------------------- > Dr. Nathan S. Watson-Haigh (publish under Haigh, N.S.) > OCE Post Doctoral Fellow > CSIRO Livestock Industries > J M Rendel Laboratory > Rockhampton > QLD 4701 Tel: +61 (0)7 4923 8121 > Australia Fax: +61 (0)7 4923 8222 > > _______________________________________________ > Bioconductor mailing list > Bioconductor at stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor >

Hi James, Since I posted, I actually found your affycoretools vignette and this was really nice to see and very well explained as I'd been doing something similar over recent weeks. One feature that would be nice would be to specify the dir that the plots are save to and/or a file prefix so if you run different normalisations, the files don't get overwritten. RE: vector not array The documentation for genefilter needs changing as it says: It returns an array of logical values (suitable for subscripting) of the same length as there are rows in 'expr'. For each row of 'expr' the returned value is 'TRUE' if the row passed all the filter functions. Cheers Nathan -----Original Message----- From: James W. MacDonald [mailto:jmacdon at med.umich.edu] Sent: Thu 19/06/2008 21:15 To: Watson-Haigh, Nathan (LI, Rockhampton) Cc: bioconductor at stat.math.ethz.ch Subject: Re: [BioC] Using genefilter Hi Nathan, Nathan.Watson-Haigh at csiro.au wrote: > I've managed to use genefilter to generate an array of FALSE and TRUE's > for the affy bovine array I'm working with: > > -- start code -- > # expSet.RNA.RMAqndata is an ExpressionSet object with log2 expression > values > > f1<-anyNA > f2<-kOverA(16, log2(100)) # filter out genes with low/no expression > over all arrays > ff<-filterfun(f1,f2) > wh<-genefilter(expSet.RNA.RMAqndata, ff) > sum(wh) # number of genes passing the filter > -- end code -- > > I was wondering if someone could shed light on how to use the wh array > downstream with things such as limma etc. > What you have isn't an array - it's a vector of TRUE/FALSE that you can use to subset your ExpressionSet to remove the reporters that don't pass the filter. expSet.RNA.RMAqndata.filtered <- expSet.RNA.RMAqndata[wh,] And then you use the filtered data with limma, etc. Best, Jim > > Thanks > Nathan > > > ------------------------------------------------------------- > Dr. Nathan S. Watson-Haigh (publish under Haigh, N.S.) > OCE Post Doctoral Fellow > CSIRO Livestock Industries > J M Rendel Laboratory > Rockhampton > QLD 4701 Tel: +61 (0)7 4923 8121 > Australia Fax: +61 (0)7 4923 8222 > > _______________________________________________ > Bioconductor mailing list > Bioconductor at stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor >

I only just came across affycoretools yesterday and haven't ready the whole vignette yet, so, bearing that in mind.... I'm using the affystart function and wanted to look at some QC plots from various normalisations. I found that after performing RMA normalisations, I had to move the files ("Density plot.pdf" and "Digestion plot.pdf") to an "rma" dir before doing gcrma normalisation to stop the files being clobbered. Allowing one to define a filename prefix e.g. "rma_" and/or specifying a dir to save the files to would be handy. While I'm at it, I also have another suggestion: Would it possible to add the option to specify the sample names to be used in the charts. Currently, I get awful CEL filenames/path as the sample names. After the eSet is created, I use "sampleNames(eSet)<-Names" and then generate the PCA plot. However, once the eSet is created, I can't generate the density or degradation plots to rectify this. Cheers Nathan -----Original Message----- From: James W. MacDonald [mailto:jmacdon@med.umich.edu] Sent: Friday, 20 June 2008 8:53 AM To: Watson-Haigh, Nathan (LI, Rockhampton) Cc: bioconductor at stat.math.ethz.ch Subject: Re: [BioC] Using genefilter Hi Nathan, Nathan.Watson-Haigh at csiro.au wrote: > Hi James, > > Since I posted, I actually found your affycoretools vignette and this > was really nice to see and very well explained as I'd been doing > something similar over recent weeks. One feature that would be nice > would be to specify the dir that the plots are save to and/or a file > prefix so if you run different normalisations, the files don't get > overwritten. To which function are you referring? I normally use affycoretools from within an Sweave document, so I can use a directive there to specify the folder to put the pdfs. Because of this it hasn't occurred to me to specify any directories. However, I am open to suggestions. Best, Jim > > RE: vector not array The documentation for genefilter needs changing > as it says: > > It returns an array of logical values (suitable for subscripting) of > the same length as there are rows in 'expr'. For each row of 'expr' > the returned value is 'TRUE' if the row passed all the filter > functions. > > Cheers Nathan > > -----Original Message----- From: James W. MacDonald > [mailto:jmacdon at med.umich.edu] Sent: Thu 19/06/2008 21:15 To: > Watson-Haigh, Nathan (LI, Rockhampton) Cc: > bioconductor at stat.math.ethz.ch Subject: Re: [BioC] Using genefilter > > Hi Nathan, > > Nathan.Watson-Haigh at csiro.au wrote: > > I've managed to use genefilter to generate an array of FALSE and > > TRUE's for the affy bovine array I'm working with: > > > > -- start code -- # expSet.RNA.RMAqndata is an ExpressionSet object > > with log2 expression values > > > > f1<-anyNA f2<-kOverA(16, log2(100)) # filter out genes with > > low/no expression over all arrays ff<-filterfun(f1,f2) > > wh<-genefilter(expSet.RNA.RMAqndata, ff) sum(wh) # number of > > genes passing the filter -- end code -- > > > > I was wondering if someone could shed light on how to use the wh > > array downstream with things such as limma etc. > > > > What you have isn't an array - it's a vector of TRUE/FALSE that you > can use to subset your ExpressionSet to remove the reporters that > don't pass the filter. > > expSet.RNA.RMAqndata.filtered <- expSet.RNA.RMAqndata[wh,] > > And then you use the filtered data with limma, etc. > > Best, > > Jim > > > > Thanks Nathan > > > > > > ------------------------------------------------------------- Dr. > > Nathan S. Watson-Haigh (publish under Haigh, N.S.) OCE Post > > Doctoral Fellow CSIRO Livestock Industries J M Rendel Laboratory > > Rockhampton QLD 4701 Tel: +61 (0)7 > > 4923 8121 Australia Fax: +61 (0)7 4923 > > 8222 > > > > _______________________________________________ Bioconductor > > mailing list Bioconductor at stat.math.ethz.ch > > https://stat.ethz.ch/mailman/listinfo/bioconductor Search the > > archives: > > http://news.gmane.org/gmane.science.biology.informatics.conductor > > > > >