Weight functions for Agilent chips (limma)

0

Entering edit mode

Malard, Joel M ▴ 100

@malard-joel-m-917

Last seen 10.2 years ago

Dear All, I was wondering what other people have been using for the weight function in read.maimages with Agilent arrays? I tried the following: wtAgilent.GFilter <- function(qta) { qta[,"gIsPosAndSignif"] } wtAgilent.RGFilter <- function(qta) { (qta[,"rIsPosAndSignif"]+qta[,"gIsPosAndSignif"])/2.0 } wtAgilent.RFilter <- function(qta) { qta[,"rIsPosAndSignif"] } wtAgilent.mRGFilter <- function(qta) { mapply(min,qta[,"gIsPosAndSignif"],qta[,"rIsPosAndSignif"]) } The last one seems to give somewhat "better results by-eye" when followed by loess normalization but is rather subjective. Best regards, Joel Malard > -----Original Message----- > From: Malard, Joel M > Sent: Friday, September 17, 2004 12:44 PM > To: 'bioconductor@stat.math.ethz.ch' > Subject: Loess normalization for Agilent chips > > > I am struggling to get data from Agilent cDNA arrays into > BioConductor. It seems to me much easier to get the data in affy's > normalize.loess() than in the other cDNA array packages. > > Given that "one who get a bargain get what he pays for", does anyone > has comments, recommendations or warnings to share about using an Affy > normalization procedure on cDNA data? > > Thanks, > > Joel M. Malard, Ph.D. > Scientist IV > Pacific Northwest National Laboratory > Battelle Boulevard, PO Box 999 > Mail Stop K1-85 > Richland, WA 99352 > > "I love the audacity of those who have everything to loose from it; > the moderation of those who have nothing to gain from it." Rostand, > Jean (1894-1977) > > [[alternative HTML version deleted]]

Normalization Normalization • 1.2k views

ADD COMMENT • link 20.2 years ago Malard, Joel M ▴ 100

0

Entering edit mode

Giovanni Coppola ▴ 250

@giovanni-coppola-893

Last seen 10.2 years ago

Hi Joel, actually, I was trying to use the 8 outlier fields (columns AZ-BG of the Agilent output file) in a way similar to the 'flag' fields of other platfoms... So, I started with mywtfun <- function(exclude.flags=c(1,2,3)) function(obj) 1-(obj$rIsBGPopnOL %in% exclude.flags) and thenRG <-read.maimages(...blabla... wt.fun=mywtfun(c(1))) and maybe there's a way to include the other fields as well.... Any other ideas? Cheers Giovanni At 02:56 PM 9/22/2004, Malard, Joel M wrote: >Dear All, > >I was wondering what other people have been using for the weight >function in read.maimages with Agilent arrays? I tried the following: > >wtAgilent.GFilter <- function(qta) { qta[,"gIsPosAndSignif"] } >wtAgilent.RGFilter <- function(qta) { >(qta[,"rIsPosAndSignif"]+qta[,"gIsPosAndSignif"])/2.0 } >wtAgilent.RFilter <- function(qta) { qta[,"rIsPosAndSignif"] } >wtAgilent.mRGFilter <- function(qta) { >mapply(min,qta[,"gIsPosAndSignif"],qta[,"rIsPosAndSignif"]) } > >The last one seems to give somewhat "better results by-eye" when >followed by loess normalization but is rather subjective. > >Best regards, > >Joel Malard > > > -----Original Message----- > > From: Malard, Joel M > > Sent: Friday, September 17, 2004 12:44 PM > > To: 'bioconductor@stat.math.ethz.ch' > > Subject: Loess normalization for Agilent chips > > > > > > I am struggling to get data from Agilent cDNA arrays into > > BioConductor. It seems to me much easier to get the data in affy's > > normalize.loess() than in the other cDNA array packages. > > > > Given that "one who get a bargain get what he pays for", does anyone > > has comments, recommendations or warnings to share about using an Affy > > normalization procedure on cDNA data? > > > > Thanks, > > > > Joel M. Malard, Ph.D. > > Scientist IV > > Pacific Northwest National Laboratory > > Battelle Boulevard, PO Box 999 > > Mail Stop K1-85 > > Richland, WA 99352 > > > > "I love the audacity of those who have everything to loose from it; > > the moderation of those who have nothing to gain from it." Rostand, > > Jean (1894-1977) > > > > > > [[alternative HTML version deleted]] > >_______________________________________________ >Bioconductor mailing list >Bioconductor@stat.math.ethz.ch >https://stat.ethz.ch/mailman/listinfo/bioconductor [[alternative HTML version deleted]]

ADD COMMENT • link 20.2 years ago Giovanni Coppola ▴ 250

0

Entering edit mode

Malard, Joel M ▴ 100

@malard-joel-m-917

Last seen 10.2 years ago

Hi Giovanni, Sorry, being new to R programs, it took me some time to parse your own. Could something of the sort be useful to you? wtAgilent.mRGOLFilter <- function(qta) { mapply(min,1-qta[,"gIsFeatNonUnifOL"],1-qta[,"gIsFeatNonUnifOL"], 1-qta[,"gIsBGNonUnifOL"],1-qta[,"gIsBGNonUnifOL"], 1-qta[,"gIsFeatPopnOL"],1-qta[,"gIsFeatPopnOL"], 1-qta[,"gIsBGPopnOL"],1-qta[,"gIsBGPopnOL"]) } I don't see (visually) much difference (for this data set) between this last filter and the previous wtAgilent.mRGFilter <- function(qta) { mapply(min,qta[,"gIsPosAndSignif"],qta[,"rIsPosAndSignif"]) } best regards, Joel P.S. For completeness sake, the Agilent manual describes 8 0/1 variables related to outliers: "gIsFeatNonUnifOL" and "rIsFeatNonUnifOL" : a value of 1 indicates Feature is a non-uniformity outlier in g(r) Boolean flag indicating if a feature is a NonUniformity Outlier or not. A feature is non-uniform if the pixel noise of feature exceeds a threshold established for a "uniform" feature. "gIsBGNonUnifOL" and " rIsBGNonUnifOL" : a value of 1 indicates Local background is a non-uniformity outlier in g(r). "gIsFeatPopnOL" and " rIsFeatPopnOL" : a value of 1 indicates Feature is a population outlier in g(r). Probes with replicate features on a microarray are examined using population statistics. A feature is a population outlier if its signal is less than a lower threshold or exceeds an upper threshold determined using the interquartile range (i.e., IQR) of the population. "gIsBGPopnOL" and "rIsBGPopnOL" : a value of 1 indicates local background is a population outlier in g(r). > Hi Joel, > actually, I was trying to use the 8 outlier fields (columns AZ-BG of the Agilent output file) in a way similar to the 'flag' fields of other platfoms... > So, I started with > mywtfun <- function(exclude.flags=c(1,2,3)) function(obj) > 1-(obj$rIsBGPopnOL %in% exclude.flags) > and thenRG <-read.maimages(...blabla... wt.fun=mywtfun(c(1))) > > and maybe there's a way to include the other fields as well.... > Any other ideas? > > Cheers > Giovanni [[alternative HTML version deleted]]

ADD COMMENT • link 20.2 years ago Malard, Joel M ▴ 100

0

Entering edit mode

Hi Joel, thanks for the function, which is useful in excluding flagged features in a way similar to the Imagene files. A simple barplot can give you an idea of the valid spots: barplot(RG$weights,ylab="OK spots") Last step, I guess you want to replace the 'g' with a 'r' in the second column: >wtAgilent.mRGOLFilter <- function(qta) { > mapply(min,1-qta[,"gIsFeatNonUnifOL"],1-qta[,"rlsFeatNonUnifOL"], > 1-qta[,"gIsBGNonUnifOL"],1-qta[,"rIsBGNonUnifOL"], > 1-qta[,"gIsFeatPopnOL"],1-qta[,"rIsFeatPopnOL"], > 1-qta[,"gIsBGPopnOL"],1-qta[,"rIsBGPopnOL"]) > } Best regards Giovanni At 05:30 PM 9/23/2004, you wrote: >Hi Giovanni, > >Sorry, being new to R programs, it took me some time to parse your own. > >Could something of the sort be useful to you? > >wtAgilent.mRGOLFilter <- function(qta) { > mapply(min,1-qta[,"gIsFeatNonUnifOL"],1-qta[,"gIsFeatNonUnifOL"], > 1-qta[,"gIsBGNonUnifOL"],1-qta[,"gIsBGNonUnifOL"], > 1-qta[,"gIsFeatPopnOL"],1-qta[,"gIsFeatPopnOL"], > 1-qta[,"gIsBGPopnOL"],1-qta[,"gIsBGPopnOL"]) > } >I don't see (visually) much difference (for this data set) between this >last filter and the previous > >wtAgilent.mRGFilter <- function(qta) { >mapply(min,qta[,"gIsPosAndSignif"],qta[,"rIsPosAndSignif"]) } > >best regards, > >Joel > >P.S. For completeness sake, the Agilent manual describes 8 0/1 variables >related to outliers: > >"gIsFeatNonUnifOL" and "rIsFeatNonUnifOL" : > a value of 1 indicates Feature is a non-uniformity outlier in g(r) > Boolean flag indicating > if a feature is a NonUniformity Outlier or not. A feature is > non-uniform if the pixel noise > of feature exceeds a threshold established for a "uniform" feature. > >"gIsBGNonUnifOL" and " rIsBGNonUnifOL" : > a value of 1 indicates Local background is a non-uniformity outlier > in g(r). > >"gIsFeatPopnOL" and " rIsFeatPopnOL" : > a value of 1 indicates Feature is a population outlier in g(r). > Probes with replicate features on > a microarray are examined using population statistics. A feature is a > population outlier if its > signal is less than a lower threshold or exceeds an upper threshold > determined using the > interquartile range (i.e., IQR) of the population. > >"gIsBGPopnOL" and "rIsBGPopnOL" : > a value of 1 indicates local background is a population outlier in g(r). > > > Hi Joel, > > actually, I was trying to use the 8 outlier fields (columns AZ-BG of > the Agilent output file) in a way similar to the 'flag' fields of other > platfoms... > > So, I started with > > mywtfun <- function(exclude.flags=c(1,2,3)) function(obj) > > 1-(obj$rIsBGPopnOL %in% exclude.flags) > > and thenRG <-read.maimages(...blabla... wt.fun=mywtfun(c(1))) > > > > and maybe there's a way to include the other fields as well.... > > Any other ideas? > > > > Cheers > > Giovanni [[alternative HTML version deleted]]

ADD REPLY • link 20.1 years ago Giovanni Coppola ▴ 250

Login before adding your answer.