Hi all,
I'm trying to get familiar with the limma package, but I've run into
something
I don't understand. After I load the package, if I type:
?normalizeWithinArrays
The documentation doesn't give a list of available normalization
methods.
Instead it refers me to:
"An overview of LIMMA functions for normalization is given in
4.Normalization."
However, when I go to the "limma: Linear Models for Microarray Data
User's
Guide" (a pdf in the limma folder of my /library) there is no section
on
normalization there. I also didn't see a vignette for limma. Since
nobody
else has asked about this, I figure the documentation is here
somewhere, and I
just don't know where to find it! Any guidance much appreciated.
Paul
limma user's Guide is 19-May-2003.
limma is 1.0.8
R is 1.7.1
OS is WinXP SP1
----------------------------------------
This mail sent through www.mywaterloo.ca
Dear Paul,
At 05:24 AM 9/08/2003, Paul Boutros wrote:
>Hi all,
>
>I'm trying to get familiar with the limma package, but I've run into
>something
>I don't understand. After I load the package, if I type:
>?normalizeWithinArrays
>The documentation doesn't give a list of available normalization
methods.
>Instead it refers me to:
>"An overview of LIMMA functions for normalization is given in
>4.Normalization."
I have reproduced below the first few paragraphs of the help document
that
one gets by typing ?normalizeWithinArrays. As you can see,
documentation
for the argument "method" gives a list of available normalization
methods.
There are further details further down in the document under
"Details". The
line that you have quoted is the second last line of the help
document. It
would seem that you are only reading the last two lines of the
document.
------------------
normalizeWithinArrays package:limma R Documentation
Normalize Within Arrays
Description:
Normalize the expression log-ratios for one or more two-colour
spotted microarray experiments so that the log-ratios average to
zero within each array or sub-array.
Usage:
normalizeWithinArrays(object, layout, method="printtiploess",
weights=object$weights, span=0.3, iterations=4, controlspots=NULL,
df=5,
robust="M")
MA.RG(object)
Arguments:
object: object of class `list', `RGList' or `MAList' containing
two-color microarray data
layout: list specifying the dimensions of the spot matrix and the
grid matrix
method: character string specifying the normalization method.
Choices
are `"none"', `"median"', `"loess"', `"printtiploess"',
`"composite"' and `"robustspline"'. A partial string
sufficient to uniquely identify the choice is permitted. If
no method is recognized, then unnormalized M and A-values
are
returned.
------------------ rest of document deleted -------------------
>However, when I go to the "limma: Linear Models for Microarray Data
User's
>Guide" (a pdf in the limma folder of my /library) there is no section
on
>normalization there. I also didn't see a vignette for limma.
The User's Guide is the vignette for LIMMA. Although there isn't a
separate
section on normalization, there are four complete data case studies
all of
which include commands for normalization.
> Since nobody
>else has asked about this, I figure the documentation is here
somewhere,
>and I
>just don't know where to find it! Any guidance much appreciated.
>
>Paul
>
>limma user's Guide is 19-May-2003.
>limma is 1.0.8
It might be a good idea to install the development version of limma
which
is up to version 1.1.10. The html version of the User's Guide is also
updated somewhat in the development version.
Regards
Gordon
>R is 1.7.1
>OS is WinXP SP1
Hi Sir,
> -----Original Message-----
> From: Gordon Smyth [mailto:smyth@wehi.edu.au]
> Sent: Saturday, August 09, 2003 4:09 AM
> To: Paul.Boutros@utoronto.ca
> Cc: bioconductor@stat.math.ethz.ch
> Subject: Re: [BioC] Limma Documentation: normalization functions
>
>
> Dear Paul,
>
> At 05:24 AM 9/08/2003, Paul Boutros wrote:
> >Hi all,
> >
> >I'm trying to get familiar with the limma package, but I've run
into
> >something
> >I don't understand. After I load the package, if I type:
> >?normalizeWithinArrays
> >The documentation doesn't give a list of available normalization
methods.
> >Instead it refers me to:
> >"An overview of LIMMA functions for normalization is given in
> >4.Normalization."
>
> I have reproduced below the first few paragraphs of the help
> document that
> one gets by typing ?normalizeWithinArrays. As you can see,
documentation
> for the argument "method" gives a list of available normalization
> methods.
> There are further details further down in the document under
> "Details". The
> line that you have quoted is the second last line of the help
> document. It
> would seem that you are only reading the last two lines of the
document.
>
> ------------------
> normalizeWithinArrays package:limma R Documentation
>
> Normalize Within Arrays
>
> Description:
>
> Normalize the expression log-ratios for one or more two-colour
> spotted microarray experiments so that the log-ratios average
to
> zero within each array or sub-array.
>
> Usage:
>
> normalizeWithinArrays(object, layout, method="printtiploess",
> weights=object$weights, span=0.3, iterations=4, controlspots=NULL,
df=5,
> robust="M")
> MA.RG(object)
>
> Arguments:
>
> object: object of class `list', `RGList' or `MAList' containing
> two-color microarray data
>
> layout: list specifying the dimensions of the spot matrix and the
> grid matrix
>
> method: character string specifying the normalization method.
Choices
> are `"none"', `"median"', `"loess"', `"printtiploess"',
> `"composite"' and `"robustspline"'. A partial string
> sufficient to uniquely identify the choice is permitted.
If
> no method is recognized, then unnormalized M and A-values
are
> returned.
> ------------------ rest of document deleted -------------------
This is weird: I get totally different output. I've appended mine
below.
Is this just a version issue, or is it something I should worry about?
------------------ Begin document ------------------------------
normalizeWithinArrays package:limma R Documentation
Normalize Within Arrays - generic
Description:
A generic function which normalizes microarray data.
Normalization
is intended to remove from the expression measures any systematic
trends which arise from the microarray technology rather than
from
differences between the probes or between the target RNA samples
hybridized to the arrays.
Usage:
normalizeWithinArrays(object,...)
Arguments:
object: a data object containing microarray data
...: any other arguments
Methods:
Methods are defined in limma for `object' an object of class
`list', `RGList' or `MAList'.
Type `showMethods("normalizeBetweenArrays")' at the R prompt for
a
complete list of classes for which methods are defined.
Author(s):
Gordon Smyth
See Also:
An overview of LIMMA functions for normalization is given in
4.Normalization.
See also `normalizeBetweenArrays' and `maNorm' in the marrayNorm
package.
------------------ End document --------------------------------
> >However, when I go to the "limma: Linear Models for Microarray
> Data User's
> >Guide" (a pdf in the limma folder of my /library) there is no
section on
> >normalization there. I also didn't see a vignette for limma.
>
> The User's Guide is the vignette for LIMMA. Although there isn't
> a separate
> section on normalization, there are four complete data case
> studies all of
> which include commands for normalization.
>
> > Since nobody
> >else has asked about this, I figure the documentation is here
somewhere,
> >and I
> >just don't know where to find it! Any guidance much appreciated.
> >
> >Paul
> >
> >limma user's Guide is 19-May-2003.
> >limma is 1.0.8
>
> It might be a good idea to install the development version of limma
which
> is up to version 1.1.10. The html version of the User's Guide is
also
> updated somewhat in the development version.
Okay, will do. Thanks!
Paul
> Regards
> Gordon
>
> >R is 1.7.1
> >OS is WinXP SP1
Dear Paul,
At 04:38 AM 10/08/2003, Paul Boutros wrote:
> > -----Original Message-----
> > From: Gordon Smyth [mailto:smyth@wehi.edu.au]
> > Sent: Saturday, August 09, 2003 4:09 AM
> > To: Paul.Boutros@utoronto.ca
> > Cc: bioconductor@stat.math.ethz.ch
> > Subject: Re: [BioC] Limma Documentation: normalization functions
> >
> >
> > Dear Paul,
> >
> > At 05:24 AM 9/08/2003, Paul Boutros wrote:
> > >Hi all,
> > >
> > >I'm trying to get familiar with the limma package, but I've run
into
> > >something
> > >I don't understand. After I load the package, if I type:
> > >?normalizeWithinArrays
> > >The documentation doesn't give a list of available normalization
methods.
> > >Instead it refers me to:
> > >"An overview of LIMMA functions for normalization is given in
> > >4.Normalization."
> >
> > I have reproduced below the first few paragraphs of the help
> > document that
> > one gets by typing ?normalizeWithinArrays. As you can see,
documentation
> > for the argument "method" gives a list of available normalization
> > methods.
> > There are further details further down in the document under
> > "Details". The
> > line that you have quoted is the second last line of the help
> > document. It
> > would seem that you are only reading the last two lines of the
document.
... help document deleted
>This is weird: I get totally different output. I've appended mine
below.
>Is this just a version issue, or is it something I should worry
about?
Ah, I can see what is happening now. My apologies.
In the BioC Release 1.2 version of limma, which was limma 1.0.8, I
wrote
normalizeWithinArrays as a new-style generic function. That means that
normalizeWithinArrays doesn't do anything itself, it just passes the
information on to more specific functions called "Methods" which
actually
do the work, depending on the type of data which you input. What you
were
supposed to do was to see that normalizeWithinArrays accepts RGList
objects
and then type ?normalizeWithinArrays.RGList for the more detailed
information. (Alternatively you could type
helpMethods("normalizeWithinArrays") which would give you a menu of
more
detailed help pages, or else use the html contents help page for
limma.)
You are far from the only one who found this too obscure. I decided
that
there is no good way to write documentation for new-style generic
functions
within the R help system as it currently stands, so I have gone back
to a
programming style in the development version of limma which does not
use
the formal object-orientated programming system. This means that the
documentation for functions like normalizeWithinArrays is now much
more
direct and you should find it more transparent.
I recommend that you get limma 1.1.10. An added advantage is that the
normalization routines are now much faster and more robust. If you
can't
get it from the Bioconductor.org www site, then you can get it from
http://bioinf.wehi.edu.au/limma/. If you have installation rights on
your
computer you may be able to type
install.packages("limma",contriburl="http://bioinf.wehi.edu.au/limma")
in R to get automatic installation of the latest version.
Best wishes
Gordon
>------------------ Begin document ------------------------------
>normalizeWithinArrays package:limma R Documentation
>
>Normalize Within Arrays - generic
>
>Description:
>
> A generic function which normalizes microarray data.
Normalization
> is intended to remove from the expression measures any
systematic
> trends which arise from the microarray technology rather than
from
> differences between the probes or between the target RNA
samples
> hybridized to the arrays.
>
>Usage:
>
> normalizeWithinArrays(object,...)
>
>Arguments:
>
> object: a data object containing microarray data
>
> ...: any other arguments
>
>Methods:
>
> Methods are defined in limma for `object' an object of class
> `list', `RGList' or `MAList'.
>
> Type `showMethods("normalizeBetweenArrays")' at the R prompt
for a
> complete list of classes for which methods are defined.
>
>Author(s):
>
> Gordon Smyth
>
>See Also:
>
> An overview of LIMMA functions for normalization is given in
> 4.Normalization.
>
> See also `normalizeBetweenArrays' and `maNorm' in the
marrayNorm
> package.
>------------------ End document --------------------------------
... earlier correspondence deleted