Yep, but plz note that for the calculation of the SF no reference array is used, but just a set value of 100 (500). This is in contrast to e.g. dChip, in which of the [experimental] arrays is used as reference, and therefore the set value varies between experiments. > -----Original Message----- > From: bioconductor-bounces at stat.math.ethz.ch > [mailto:bioconductor-bounces at stat.math.ethz.ch] On Behalf Of > James Anderson > Sent: 29 June 2007 17:40 > To: Hooiveld, Guido; bioconductor > Subject: Re: [BioC] question about scale factor for Affy chip > > Thanks for your email. So the criterion "SFs should be within > 3 fold" actually means that the ratio of the scale factor of > each array to the scale factor of the reference (mean) array > should be within 3 fold. > > "Hooiveld, Guido" <guido.hooiveld at="" wur.nl=""> wrote: > It is A, since you scale all arrays to the same target value > (100 is default in BioC, 500 is default in GCOS). > > According to Affymetrix'criteria the difference between the > SFs should not vary more than a factor of 3. Since in your > case the SFs are within the same range (10.-10.2), the SF > does not indicate a problem. If you would have had 10.2. > 10.1, 10.4, 40.9 and 2.7, the SF indiactes the last two > arrays deviate from the other arrays. > > HTH, > Guido > > > > -----Original Message----- > > From: bioconductor-bounces at stat.math.ethz.ch > > [mailto:bioconductor-bounces at stat.math.ethz.ch] On Behalf Of James > > Anderson > > Sent: 29 June 2007 16:32 > > To: bioconductor > > Subject: [BioC] question about scale factor for Affy chip > > > > Hi, > > > > I am using the R package SimpleAffy, now I have a > conceptual question > > regarding the scale factor. Seems that Affymetrix recommend > the scale > > factor should be within 3 fold, the way I understand the > scale factor > > is that it is the ratio between the trimmed mean of one chip > > (normalized by MAS5) to the target value (say, 100 or 500). > According > > to the manual for QC in SimpleAffy: > > > > "Large variations in scale factors signal cases where the > > normalisation assumptions are likely to fail due to issues > with sample > > quality or amount of starting material." > > > > I am getting a little confused about EXACTLY what does scale factor > > mean: > > A. The ratio between the trimmed mean of each chip to a > target value > > B. The ratio between the trimmed mean of each chip to the > trimmed mean > > of all the chips. > > > > because the above section seems to me that as long as there > is small > > variation in the scale factor, the quality is no problem. > What about > > all the arrays have scale factor in the range (10.1-10.2)? > The scale > > factor has almost no variation, but they are much higher > than 3. or I > > am misunderstanding some part? > > > > > > Thanks a lot! > > > > James > > > > > > --------------------------------- > > Pinpoint customers who are looking for what you sell. > > [[alternative HTML version deleted]] > > > > _______________________________________________ > > 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 > > > > > > --------------------------------- > > [[alternative HTML version deleted]] > > _______________________________________________ > 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 >