Hi,
I have just been handed a project that looks like the experimental
design has not been ideal and I was wondering if anyone had any ideas
how I could get something out of it.
Here is the design:
3 microarrays
Array 1: A vs B
Array 2: B vs C
Array 3: C vs A
Where A, B, C are different samples.
The question that the experimenter would like to ask is what are the
similarities between the three samples and what are the differences.
Has anyone got any bright ideas how to proceed? Without a control I
can't really see how to do this.
Thanks
Dan
--
**************************************************************
Daniel Brewer, Ph.D.
Institute of Cancer Research
Email: daniel.brewer at icr.ac.uk
**************************************************************
The Institute of Cancer Research: Royal Cancer Hospital, a charitable
Company Limited by Guarantee, Registered in England under Company No.
534147 with its Registered Office at 123 Old Brompton Road, London SW7
3RP.
This e-mail message is confidential and for use by the
addre...{{dropped}}
Hi Daniel,
Daniel Brewer wrote:
> Hi,
>
> I have just been handed a project that looks like the experimental
> design has not been ideal and I was wondering if anyone had any
ideas
> how I could get something out of it.
>
> Here is the design:
> 3 microarrays
> Array 1: A vs B
> Array 2: B vs C
> Array 3: C vs A
>
> Where A, B, C are different samples.
Back when I was doing two color work, this was considered the height
of
good experimental design (the reference design was disparaged because
you were wasting half of your time/effort on samples with no intrinsic
value). Maybe things have changed?
Anyway, the maanova package will certainly handle this sort of
analysis.
A pubmed search for Kerr and Churchill will result in lots of hits for
the papers they wrote describing the model they use to fit these data.
The maanova package also has a vignett in which they analyze an
experiment that is sort of similar.
I don't see why you couldn't use limma either. I don't know offhand
what
the design matrix would look like (as I mentioned, it's been years
since
I did two color stuff), but I am sure there is something in the limma
User's Guide that you could use to figure things out.
Best,
Jim
>
> The question that the experimenter would like to ask is what are the
> similarities between the three samples and what are the differences.
> Has anyone got any bright ideas how to proceed? Without a control I
> can't really see how to do this.
>
> Thanks
>
> Dan
--
James W. MacDonald
University of Michigan
Affymetrix and cDNA Microarray Core
1500 E Medical Center Drive
Ann Arbor MI 48109
734-647-5623
The simplest way to handle this in limma is to use a single channel
analysis, using block=array. The method is described clearly in the
vignette. MAANOVA will also handle this, although it might be a bit
harder to set up.
There are several papers showing that in theory this is the most
efficient design for 3 conditions with 2 channel arrays. There are
also several papers that confirm this experimentally. The design is
called a loop design. For 3 conditions, the loop design is a
balanced incomplete block design.
--Naomi
At 02:56 PM 9/6/2007, James W. MacDonald wrote:
>Hi Daniel,
>Daniel Brewer wrote:
> > Hi,
> >
> > I have just been handed a project that looks like the experimental
> > design has not been ideal and I was wondering if anyone had any
ideas
> > how I could get something out of it.
> >
> > Here is the design:
> > 3 microarrays
> > Array 1: A vs B
> > Array 2: B vs C
> > Array 3: C vs A
> >
> > Where A, B, C are different samples.
>
>Back when I was doing two color work, this was considered the height
of
>good experimental design (the reference design was disparaged because
>you were wasting half of your time/effort on samples with no
intrinsic
>value). Maybe things have changed?
>
>Anyway, the maanova package will certainly handle this sort of
analysis.
>A pubmed search for Kerr and Churchill will result in lots of hits
for
>the papers they wrote describing the model they use to fit these
data.
>The maanova package also has a vignett in which they analyze an
>experiment that is sort of similar.
>
>I don't see why you couldn't use limma either. I don't know offhand
what
>the design matrix would look like (as I mentioned, it's been years
since
>I did two color stuff), but I am sure there is something in the limma
>User's Guide that you could use to figure things out.
>
>Best,
>
>Jim
>
>
>
> >
> > The question that the experimenter would like to ask is what are
the
> > similarities between the three samples and what are the
differences.
> > Has anyone got any bright ideas how to proceed? Without a control
I
> > can't really see how to do this.
> >
> > Thanks
> >
> > Dan
>
>
>--
>James W. MacDonald
>University of Michigan
>Affymetrix and cDNA Microarray Core
>1500 E Medical Center Drive
>Ann Arbor MI 48109
>734-647-5623
>
>_______________________________________________
>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
Naomi S. Altman 814-865-3791 (voice)
Associate Professor
Dept. of Statistics 814-863-7114 (fax)
Penn State University 814-865-1348
(Statistics)
University Park, PA 16802-2111