On Monday 11 December 2006 09:33, Ingrid H. G. ?stensen wrote:
> Hi
>
> I am currently working on data from the Affymetrix platform. I have
gotten
> 36 arrays that are divided into 3 groups: 3 different cell lines
where two
> are derived/changed versions of the first. These 3 groups are
divided into
> 4 time points: 0 (control), and 3, 24 and 72 hours. There are 7
comparisons
> that are to be done: One for each group (3) and one for each time
point
> (4).
>
> Because this is the Human Genome U133 Plus 2 Array from Affymetrix I
can
> not read all the files into R/Bioconductor at once, I run out of
memory. I
> can divide the files into the different comparisons groups and work
on
> these but it is one thing that troubles me. If I do the seven
different
> comparisons (everything separate: loading the data, normalization,
analysis
> etc) can I compare the different results? For example: Lock at time
point
> 24 and 72 to find the differences between them, or cell group 1 and
2?
Hi, Ingrid. You might start by letting us know how you are trying to
load the
arrays into Bioconductor. There may be a way to load all of them (by
using
justRMA instead of RMA, for example). Include the output of
sessionInfo() so
that we know the OS and R version, as well as the amount of RAM in
your
machine. If you can load all the arrays simultaneously, it may make
your
life easier down the road.
Sean
Ingrid H. G. ?stensen wrote:
> Hi
>
> I am currently working on data from the Affymetrix platform. I have
> gotten 36 arrays that are divided into 3 groups: 3 different cell
> lines where two are derived/changed versions of the first. These 3
> groups are divided into 4 time points: 0 (control), and 3, 24 and 72
> hours. There are 7 comparisons that are to be done: One for each
> group (3) and one for each time point (4).
>
> Because this is the Human Genome U133 Plus 2 Array from Affymetrix I
> can not read all the files into R/Bioconductor at once, I run out of
> memory. I can divide the files into the different comparisons groups
> and work on these but it is one thing that troubles me. If I do the
> seven different comparisons (everything separate: loading the data,
> normalization, analysis etc) can I compare the different results?
For
> example: Lock at time point 24 and 72 to find the differences
between
> them, or cell group 1 and 2?
Doing everything separate like that is a very bad idea. You will not
be
able to tell if differential expression is due to biological
differences
or because you ran the analysis separately.
Instead, you should try either justRMA(), which is more memory
efficient
and may be able to process all the chips at once, or RMAexpress, which
is a standalone executable that can run lots of chips with not much
memory. You can find RMAexpress here:
http://rmaexpress.bmbolstad.com/
Best,
Jim
>
> Regards, Ingrid
>
>
>
> --- Ingrid H.G. ?stensen Bioinformatics user support (NMC)
Department
> of Tumor Biology (Montebello) Rikshospitalet-Radiumhospitalet
Medical
> Center 0310 Oslo
>
> E -mail: ingrid at microarray.no Phone: +47 22 93 54 11 Fax: +47
22 52
> 24 21
>
>
>
> [[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
--
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.
On Tuesday 12 December 2006 05:14, Ingrid H. G. ?stensen wrote:
> Hi
>
> More information:
> 1,50 GB RAM
>
> > sessionInfo()
>
> R version 2.4.0 (2006-10-03)
> i386-pc-mingw32
>
> locale:
> LC_COLLATE=Norwegian (Bokm?l)_Norway.1252;LC_CTYPE=Norwegian
> (Bokm?l)_Norway.1252;LC_MONETARY=Norwegian
> (Bokm?l)_Norway.1252;LC_NUMERIC=C;LC_TIME=Norwegian
(Bokm?l)_Norway.1252
>
> attached base packages:
> [1] "methods" "stats" "graphics" "grDevices" "utils"
"datasets"
> [7] "base"
>
> From the runing of the script:
> > library(limma)
> > library(affyQCReport)
>
> Loading required package: Biobase
> Loading required package: tools
>
> Welcome to Bioconductor
>
> Vignettes contain introductory material. To view, type
> 'openVignette()' or start with 'help(Biobase)'. For details
> on reading vignettes, see the openVignette help page.
>
> Loading required package: affy
> Loading required package: affyio
> Loading required package: simpleaffy
> Loading required package: genefilter
> Loading required package: survival
> Loading required package: splines
> Welcome to 'affyQCReport' V 1.6-1
> Further information available at: www.bifix.org
> mailto: craig.parman at bifix.org
>
> > library(RColorBrewer)
> > library(xtable)
> > library(simpleaffy)
> > memory.limit(4000)
>
> NULL
>
> > dataSet <- read.affy()
Thanks, Ingrid, for the details. As Jim already suggested, using
justRMA
should probably allow you to load all of your arrays. You will get
back an
ExpressionSet which you can then use for further analyses.
Normalizing and
analyzing all your data simultaneously has significant advantages over
the
piecemeal approach.
Sean
On Wednesday 13 December 2006 02:56, Ingrid H. G. ?stensen wrote:
> Hi
>
> I usually do a quality control of the data before I turn the data
set into
> an ExpressionSet. If I use justRMA, how can I then quality control
the data
> to see if any array should be removed? Will it not be difficult if
all the
> arrays are corrected as they are read into the program?
I don't know what you mean by quality control (specifically), but for
many of
the quality-control measures, they are done independently on each
chip, so
you could still load in batches to do your QC.
Sean
If I have so many arrays that I need to use
justRMA to normalize, I do the QC in small
batches first. Using cross-array comparisons for
QC makes most sense when you expect the arrays to
be similar, so I read in all the replicates for
one condition. But for statistical differential
expression analysis, the arrays should be normalized all together.
--Naomi
At 02:56 AM 12/13/2006, Ingrid H. G. ?stensen wrote:
>Hi
>
>I usually do a quality control of the data
>before I turn the data set into an
>ExpressionSet. If I use justRMA, how can I then
>quality control the data to see if any array
>should be removed? Will it not be difficult if
>all the arrays are corrected as they are read into the program?
>
>Regards,
>Ingrid
>
>
>
>-----Original Message-----
>From: Sean Davis [mailto:sdavis2 at mail.nih.gov]
>Sent: Tue 12/12/2006 12:39
>To: Ingrid H. G. ?stensen
>Cc: bioconductor at stat.math.ethz.ch
>Subject: Re: [BioC] Compare results from different analysis
(Affymetrix)
>
>On Tuesday 12 December 2006 05:14, Ingrid H. G. ?stensen wrote:
> > Hi
> >
> > More information:
> > 1,50 GB RAM
> >
> > > sessionInfo()
> >
> > R version 2.4.0 (2006-10-03)
> > i386-pc-mingw32
> >
> > locale:
> > LC_COLLATE=Norwegian (Bokm?l)_Norway.1252;LC_CTYPE=Norwegian
> > (Bokm?l)_Norway.1252;LC_MONETARY=Norwegian
> > (Bokm?l)_Norway.1252;LC_NUMERIC=C;LC_TIME=Norwegian
(Bokm?l)_Norway.1252
> >
> > attached base packages:
> > [1] "methods" "stats" "graphics" "grDevices" "utils"
"datasets"
> > [7] "base"
> >
> > From the runing of the script:
> > > library(limma)
> > > library(affyQCReport)
> >
> > Loading required package: Biobase
> > Loading required package: tools
> >
> > Welcome to Bioconductor
> >
> > Vignettes contain introductory material. To view, type
> > 'openVignette()' or start with 'help(Biobase)'. For details
> > on reading vignettes, see the openVignette help page.
> >
> > Loading required package: affy
> > Loading required package: affyio
> > Loading required package: simpleaffy
> > Loading required package: genefilter
> > Loading required package: survival
> > Loading required package: splines
> > Welcome to 'affyQCReport' V 1.6-1
> > Further information available at: www.bifix.org
> > mailto: craig.parman at bifix.org
> >
> > > library(RColorBrewer)
> > > library(xtable)
> > > library(simpleaffy)
> > > memory.limit(4000)
> >
> > NULL
> >
> > > dataSet <- read.affy()
>
>Thanks, Ingrid, for the details. As Jim already suggested, using
justRMA
>should probably allow you to load all of your arrays. You will get
back an
>ExpressionSet which you can then use for further analyses.
Normalizing and
>analyzing all your data simultaneously has significant advantages
over the
>piecemeal approach.
>
>Sean
>
>
> [[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
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
Sean Davis <sdavis2 at="" ...=""> writes:
>
> On Wednesday 13 December 2006 02:56, Ingrid H. G. ?stensen wrote:
> > Hi
> >
> > I usually do a quality control of the data before I turn the data
set into
> > an ExpressionSet. If I use justRMA, how can I then quality control
the data
> > to see if any array should be removed? Will it not be difficult if
all the
> > arrays are corrected as they are read into the program?
>
> I don't know what you mean by quality control (specifically), but
for many of
> the quality-control measures, they are done independently on each
chip, so
> you could still load in batches to do your QC.
>
> Sean
The package affyPLM uses a multichip approach to quality assessment.
To avoid
having to do this in batches, you might try the web-based interface
RACE,
available at http://race.unil.ch
Best regards,
Darlene
--
Darlene Goldstein
?cole Polytechnique F?d?rale de Lausanne (EPFL)
Institut de math?matiques
B?timent MA, Station 8 Tel: +41 21 693 5565
CH-1015 Lausanne Fax: +41 21 693 4303
SWITZERLAND