[BioC] flagging spots in marray
Gordon K Smyth
smyth at wehi.EDU.AU
Wed Jan 3 13:40:03 CET 2007
Dear Jay,
limma functions such a normalizeWithinArrays() and lmFit() all use spot weights automatically.
As far as I know, maNorm() doesn't use spot weights.
Best wishes
Gordon
> Date: Tue, 2 Jan 2007 14:51:17 -0700
> From: Jay Konieczka <jayk at u.arizona.edu>
> Subject: [BioC] flagging spots in marray
> To: bioconductor at stat.math.ethz.ch
> Message-ID: <4C50AEE2-EBF6-4133-AF72-F6BE35FBE78E at u.arizona.edu>
> Content-Type: text/plain
>
> Hi,
>
> I'm using the marray package to analyse my 2-color cDNA array data,
> and I'm trying to sort out how to tell the various functions which
> spots to use. I understand how to set the maSub vector in the layout
> object, and my understanding is that the normalization functions will
> then use this vector to know which spots to include in the
> normalization procedures so long as subset=TRUE when I call a given
> normalization function. But how do I flag spots within specific
> arrays such that those spots will not be used in the normalization or
> in the limma methods? For example, here is some code a friend of
> mine was using to flag spots for normalization methods using the
> limma function read.maimages():
>
> #this function flags out bad spots defined as saturated, too small or
> large, absent or manually flagged as bad.
> filters <- function(x) {
> okred <- x[,"F635 Mean"] < 65534
> okgreen <- x[,"F532 Mean"] < 65534
> fat <- x[,"Dia."] > 50
> skinny <- x[,"Dia."] < 150
> present <- x[,"Normalize"] < 1
> good <- x[,"Flags"] > -50.5
> as.numeric(okred & okgreen & fat & skinny & present & good)
> }
>
> #read in data and gal file
> RG = read.maimages(targets$FileName, source="genepix",
> wt.fun=filters, other.columns=c("F635 SD", "B635 SD", "F532 SD",
> "B532 SD", "F Pixels", "B Pixels", "B532 Mean","B635 Mean"))
>
> So now, because of the wt.fun parameter, those spots are considered
> (or not) according to the filter function. What would be the
> corresponding way to do this using marray? I'm assuming I must set
> the name.W parameter when I call read.marrayRaw() to point to a
> vector indicating the weights of each spot. If so, I can handle
> this, but how do I ensure that the normalization methods will use
> this vector? If the name.W parameter is set in the marrayRaw object,
> will maNorm() know to use it by default? Or must I tell it that it
> is set in some way? What about lmFit()?
>
> Thanks,
>
> Jay
>
> --
> Jay H. Konieczka
> Ph.D. Student, Antin Lab
> Molecular & Cellular Biology
> University of Arizona
>
> Phone: 1.520.591.3446
>
> 1656 E. Mabel, MRB 317
> Tucson, AZ 85724 USA
> _____________________
>
>
>
> [[alternative HTML version deleted]]
>
>
>
> ------------------------------
>
> Message: 5
> Date: Wed, 03 Jan 2007 09:44:23 +1100
> From: Gordon Smyth <smyth at wehi.EDU.AU>
> Subject: Re: [BioC] Plate or print-order effects - limma
> To: Jenny Drnevich <drnevich at uiuc.edu>
> Cc: bioconductor at stat.math.ethz.ch, J.delasHeras at ed.ac.uk,
> Joao.Fadista at agrsci.dk
> Message-ID: <6.2.5.6.1.20070103092525.02171590 at wehi.edu.au>
> Content-Type: text/plain; charset="us-ascii"; format=flowed
>
> At 07:38 AM 3/01/2007, Jenny Drnevich wrote:
>>Hi Gordon,
>>
>>FYI - I was playing around with plotting print order effects, and
>>found a discrepancy between the help documentation and the actual
>>behavior of 'normalizeForPrintorder.rg '. The help file states "The
>>function plotPrintorder or normalizeForPrintorder.rg with plot=TRUE
>>returns no value but produces a plot as a side-effect." However,
>>when I used 'normalizeForPrintorder.rg' with plot=TRUE, it not only
>>produced the plot but spit out the entire list!
>
> You're right. I'll fix.
>
>>I'm using this function instead of the wrapper 'plotPrintorder'
>>because I'm having trouble making the necessary layout list. I could
>>only find a little bit of information about the layout list in
>>'printorder' help page - is each component of the list (ngrid.r,
>>ngrid.c, nspot.r and nspot.c) just a single number? In my case our
>>printer has 48 tips and prints 12 x 4 blocks of 420 spots, one dip
>>per spot, so would my layout just be a list of ngrid.r = 12, ngrid.c
>>= 4, nspot.r = 420, and nspot.c = 1?
>
> You also need to know the number of rows (nspot.r) and columns
> (nspot.c) of spots in each block. nspot.c is almost certainly greater than 1.
>
> You also need to check with your microarray printing unit whether
> printing of your arrays begins in the top right or top left of each
> block, and whether it proceeds by rows or columns. The default for
> normalizeForPrintorder assumes topleft by rows, which the usual setup
> at UCSF. But the AGRF in Australia uses topright by rows.
>
> See help("PrinterLayout-class"). The read.maimages() function
> produces the printer layout information automatically as RG$printer
> for some input sources including GenePix and ImageGene.
>
> Best wishes
> Gordon
>
>>Thanks,
>>Jenny
>>
>> > sessionInfo()
>>R version 2.4.0 (2006-10-03)
>>i386-pc-mingw32
>>
>>locale:
>>LC_COLLATE=English_United States.1252;LC_CTYPE=English_United
>>States.1252;LC_MONETARY=English_United
>>States.1252;LC_NUMERIC=C;LC_TIME=English_United States.1252
>>
>>attached base packages:
>>[1] "splines" "tools" "methods" "stats" "graphics" "grDevices"
>>[7] "utils" "datasets" "base"
>>
>>other attached packages:
>> affyQCReport simpleaffy made4 scatterplot3d ade4
>> "1.12.0" "2.8.0" "1.8.0" "0.3-24" "1.4-2"
>> affyPLM affydata affycoretools annaffy xtable
>> "1.10.0" "1.10.0" "1.7.5" "1.6.0" "1.4-2"
>> gcrma matchprobes biomaRt RCurl XML
>> "2.6.0" "1.6.0" "1.8.0" "0.7-0" "1.2-0"
>> GOstats Category genefilter survival KEGG
>> "2.0.3" "2.0.3" "1.12.0" "2.29" "1.14.1"
>> RBGL annotate GO graph limma
>> "1.10.0" "1.12.0" "1.14.1" "1.12.0" "2.9.1"
>> affy affyio Biobase RWinEdt
>> "1.12.1" "1.2.0" "1.12.2" "1.7-5"
>>
>>Jenny Drnevich, Ph.D.
>>
>>Functional Genomics Bioinformatics Specialist
>>W.M. Keck Center for Comparative and Functional Genomics
>>Roy J. Carver Biotechnology Center
>>University of Illinois, Urbana-Champaign
>>
>>330 ERML
>>1201 W. Gregory Dr.
>>Urbana, IL 61801
>>USA
>>
>>ph: 217-244-7355
>>fax: 217-265-5066
>>e-mail: drnevich at uiuc.edu
>
>
>
> ------------------------------
>
> Message: 6
> Date: Wed, 03 Jan 2007 10:55:27 +0000
> From: Daniel Brewer <daniel.brewer at icr.ac.uk>
> Subject: [BioC] Import GEO Soft format
> To: bioconductor at stat.math.ethz.ch
> Message-ID: <459B8B9F.4060105 at icr.ac.uk>
> Content-Type: text/plain; charset=ISO-8859-1
>
> Hi,
>
> I have got some aCGH data from GEO, what is the best way to get this
> into R in a suitable object. Is there a package that interprets Soft
> format or is the approach to produce tab-delimited files?
>
> Thanks
>
> --
> **************************************************************
>
> Daniel Brewer, Ph.D.
>
> Institute of Cancer Research
> Molecular Carcinogenesis
> MUCRC
> 15 Cotswold Road
> Sutton, Surrey SM2 5NG
> United Kingdom
>
> Tel: +44 (0) 20 8722 4109
> Fax: +44 (0) 20 8722 4141
>
> Email: daniel.brewer at icr.ac.uk
>
>
>
> ------------------------------
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
>
>
> End of Bioconductor Digest, Vol 47, Issue 2
> *******************************************
>
More information about the Bioconductor
mailing list