[Bioc-devel] ExpressionSet for qRT-PCR
mtmorgan at fhcrc.org
Mon Jan 15 18:26:12 CET 2007
"Matthias Kohl" <Matthias.Kohl at stamats.de> writes:
> Hi Jim,
> we work with 96 well plates (on the iQ5 platform of Bio-Rad).
> The next steps I'm going to take in the near future are:
> - add a function to read in the data for our platform
> - use S4 class (ExpressionSet?) to save/hold the data
If the data is really quite different from that currently stored in
ExpressionSet (i.e., gene expression data from microarray experiments)
then probably you want to create a new class for the data. This will
help associate appropriate methods with appropriate data.
This does NOT mean starting a new class, say qtPCRSet, from
scratch. If qtPCRSet is a specialized version of ExpressionSet data
(you'd like functionality that is unique to qtPCR, in addition to the
analytic methods available to ExpressionSet) then you would subclass
It seems that a more likely starting point is that you want to have
the overall coordination of assay data + phenotypic data + feature
data + experiment data + annotation, but the data analysis methods
relevant to qtPCRSet are different from those for ExpressionSet. Then
the starting point is eSet
> setClass("qtPCRSet", contains="eSet")
qtPCRSet (storageMode: lockedEnvironment)
varLabels and varMetadata: none
varLabels and varMetadata: none
experimentData: use 'experimentData(object)'
and you'll have lots of functionality (e.g., coordinated subsetting,
display, access to slots via phenoData(), etc) already available (the
'initialize' method might be tricky to get exactly right, but
ExpressionSet provides a model). More info is in the vignette "Notes
for eSet developers".
> - Implement further normalization strategies (with methods for "ExpressionSet", "matrix", "data.frame"?)
> - add plot methods
> So, the next versions of my package will mainly contain normalization strategies.
> Best regards,
> ----- original message --------
> Subject: Re: [Bioc-devel] ExpressionSet for qRT-PCR
> Sent: Sat, 13 Jan 2007
> From: James W. MacDonald<jmacdon at med.umich.edu>
>> Hi Matthias,
>> Matthias Kohl wrote:
>> > Hi Jim,
>> > I'm very interested in your work on qRT-PCR data.
>> > I also work on a package to analyze qRT-PCR data. The first version of
>> > this package includes the methods described in Vandesompele et al (2002)
>> > (cf. http://www.gene-quantification.com/vandesompeles-2002.pdf). I want
>> > to add further methods for analyzing qRT-PCR data in near future and are
>> > also thinking about using an S4 class (if it works "ExpressionSet" or a
>> > derived class) to keep all data.
>> > I submitted the first version of this package (SLqPCR) to
>> > http://bioconductor.org/uploadPackage/
>> > about two weeks ago, but did not get any feed-back so far.
>> I have been looking at your package a bit, but have not had the time yet
>> to give a reasonable review.
>> > Which methods do you want to add to your package?
>> Probably nothing fancy. I like the idea of having some quantitative way
>> to decide which control gene to use for normalization (as implemented in
>> your SLqPCR package), but for differential expression I was only
>> thinking about delta-deltaCT when there is no replication, and simple
>> t-tests (or possibly using limma's empirical Bayesian variance adjusted
>> t-tests) when there is replication.
>> I'm not yet sure I want to create an actual package for release. My core
>> has started offering SuperArrays to our clients, and if I am going to be
>> seeing a lot of them, I wanted to have a consistent framework in place
>> for the analyses.
>> What sort of analyses were you planning on adding to SLqPCR? Are you
>> intending this package for the analysis of arbitrary PCR data, or a
>> particular platform?
>> > Best regards
>> > Matthias
>> > James MacDonald schrieb:
>> >> I'm thinking about writing some functions to analyze qRT-PCR data,
>> >> specifically the SuperArrays, which come in 96 or 384 well plates. I am
>> >> thinking that an ExpressionSet would be a nice container for these data,
>> >> and I hoped to get some advice.
>> >> The data I would want to put in the ExpressionSet would consist of the
>> >> cycle threshold values (numeric), which of course would go in the exprs
>> >> slot. SuperArray also supply a file that is essentially a 96 row matrix
>> >> that has the well, the gene symbol, the UniGene ID, Entrez Gene ID, and
>> >> the gene name. Ideally I would also like to stick these data in the
>> >> ExpressionSet as well, but I am not sure where. These data are part
>> >> annotation, and part location information. Since they map the genes to
>> >> the wells, I would like to keep them in the ExpressionSet (while
>> >> annotation data are supposed to be in an external package).
>> >> Is the featureData slot a good place? I can get it to go into the data
>> >> slot of an AnnotatedDataFrame, but not the varMetaData slot (which seems
>> >> like a more logical place).
>> >> Any suggestions?
>> >> Best,
>> >> Jim
>> >> James W. MacDonald, M.S.
>> >> Biostatistician
>> >> Affymetrix and cDNA Microarray Core
>> >> University of Michigan Cancer Center
>> >> 1500 E. Medical Center Drive
>> >> 7410 CCGC
>> >> Ann Arbor MI 48109
>> >> 734-647-5623
>> >> **********************************************************
>> >> Electronic Mail is not secure, may not be read every day, and should
>> >> not be used for urgent or sensitive issues.
>> >> _______________________________________________
>> >> Bioc-devel at stat.math.ethz.ch mailing list
>> >> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>> James W. MacDonald
>> University of Michigan
>> Affymetrix and cDNA Microarray Core
>> 1500 E Medical Center Drive
>> Ann Arbor MI 48109
>> Electronic Mail is not secure, may not be read every day, and should not be
>> used for urgent or sensitive issues.
> --- original message end ----
> Bioc-devel at stat.math.ethz.ch mailing list
Martin T. Morgan
Bioconductor / Computational Biology
More information about the Bioc-devel