Dear all,
sorry to bother again but I have been following the examples from the
com.braju.sma manual with some of my data and have seen some warnings
that I would like to understand, if anyone is so kind to explain me.
I am using QuantArray .txt files for two colour chips.
1.- I wanted to extract the signal from the chips without doing
background subtraction and therefore wrote the following
> ma<-getSignal(raw, bg.subtract=FALSE)
Warning message:
Argument 'bg.subtract' in getSignal() (class RawData) is deprecated.
Please use 'bgSubtract' instead. in: getSignal.RawData(raw,
bg.subtract = FALSE)
but it still performs the calculation (as you can see below). Should I
use bgSubtract instead as stated on the message?
> ma
[1] "MAData: M (12288x6), A (12288x6), Layout: Grids: 12x4 (=48),
spots in grids: 16x16 (=256), total number of spots: 12288. Spot id's
are specified."
2.- Also, I wanted to have a look at the A values, but instead got the
following answer
> ma$A
[1] "SpotSlideArray: 2003-06-24 18:01:02"
What should I type?
3.- Finally, I normalised the chips using Lowess and got another
warning:
> normalizeWithinSlide(ma, method="l")
Warning messages:
1: Collapsing to unique x values in: approx(line, xout = A[ok])
2: Collapsing to unique x values in: approx(line, xout = A[ok])
3: Collapsing to unique x values in: approx(line, xout = A[ok])
4: Collapsing to unique x values in: approx(line, xout = A[ok])
5: Collapsing to unique x values in: approx(line, xout = A[ok])
6: Collapsing to unique x values in: approx(line, xout = A[ok])
but again the calculation is performed. What does this mean?
Thanks ever so much for your help.
David
Dear David,
com.braju.sma is Henrik Bengtsson's very interesting package. Although
some
of us work with Henrik on microarray problems, the com.braju.sma
package is
not associated with bioconductor and isn't designed to work with
bioconductor functions. You need to send any questions you have to
Henrik
rather than to us!
Best wishes
Gordon
At 03:07 AM 25/06/2003, kfbargad@lg.ehu.es wrote:
>Dear all,
>
>sorry to bother again but I have been following the examples from the
>com.braju.sma manual with some of my data and have seen some warnings
>that I would like to understand, if anyone is so kind to explain me.
>
>I am using QuantArray .txt files for two colour chips.
>
>1.- I wanted to extract the signal from the chips without doing
>background subtraction and therefore wrote the following
>
> > ma<-getSignal(raw, bg.subtract=FALSE)
>Warning message:
>Argument 'bg.subtract' in getSignal() (class RawData) is deprecated.
>Please use 'bgSubtract' instead. in: getSignal.RawData(raw,
>bg.subtract = FALSE)
>
>but it still performs the calculation (as you can see below). Should
I
>use bgSubtract instead as stated on the message?
> > ma
>[1] "MAData: M (12288x6), A (12288x6), Layout: Grids: 12x4 (=48),
>spots in grids: 16x16 (=256), total number of spots: 12288. Spot id's
>are specified."
>
>2.- Also, I wanted to have a look at the A values, but instead got
the
>following answer
> > ma$A
>[1] "SpotSlideArray: 2003-06-24 18:01:02"
>
>What should I type?
>
>3.- Finally, I normalised the chips using Lowess and got another
>warning:
>
> > normalizeWithinSlide(ma, method="l")
>Warning messages:
>1: Collapsing to unique x values in: approx(line, xout = A[ok])
>2: Collapsing to unique x values in: approx(line, xout = A[ok])
>3: Collapsing to unique x values in: approx(line, xout = A[ok])
>4: Collapsing to unique x values in: approx(line, xout = A[ok])
>5: Collapsing to unique x values in: approx(line, xout = A[ok])
>6: Collapsing to unique x values in: approx(line, xout = A[ok])
>
>but again the calculation is performed. What does this mean?
>
>Thanks ever so much for your help.
>
>David
Justin
http://naturalvariation.org
We are having a little debate about whether RMA is using MM oligos at
all.
I know that PM.only is the method that calls the gene expression
index, but
Are MM intensities used in the calculation for the background
correction? I
thought I'd put this out to get the definitive answer.
One that note, if they are used does it really help that much? If
people
are interested, I for 1 would be happy to hear thoughts (on or off
line) on
whether MM oligos are worth the space for custom affy array designs.
With
the analysis we are doing now (RMA and probe level models) it seems to
me
the answer is no.
Justin
http://naturalvariation.org
The current implementations of RMA do not use MM in any way (as you
might calculate using rma() or expresso() or justRMA()). Some of the
earlier versions used MM in the computation of the background, but
this
is no longer the case.
Thanks,
Ben
> We are having a little debate about whether RMA is using MM oligos
at all.
> I know that PM.only is the method that calls the gene expression
index, but
> Are MM intensities used in the calculation for the background
correction? I
> thought I'd put this out to get the definitive answer.
>
--
Ben Bolstad <bolstad@stat.berkeley.edu>
http://www.stat.berkeley.edu/~bolstad
i want to add to ben's answer that the difference between the RMA
using MM for background and the current version is small.
Some collegues and I are working on a new RMA that uses MM for
non-specific binding correction in such a way that it improves RMAs
accuracy a bit without lossing precision. By the end of june ill make
this
work public. once you see the gains in accuracy you can decide if they
are
worth the space.
On Tue, 24 Jun 2003, Ben Bolstad wrote:
> The current implementations of RMA do not use MM in any way (as you
> might calculate using rma() or expresso() or justRMA()). Some of the
> earlier versions used MM in the computation of the background, but
this
> is no longer the case.
>
> Thanks,
>
> Ben
>
> > We are having a little debate about whether RMA is using MM oligos
at all.
> > I know that PM.only is the method that calls the gene expression
index, but
> > Are MM intensities used in the calculation for the background
correction? I
> > thought I'd put this out to get the definitive answer.
> >
>
> --
> Ben Bolstad <bolstad@stat.berkeley.edu>
> http://www.stat.berkeley.edu/~bolstad
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor@stat.math.ethz.ch
> https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor
>
In affy the plotLocation function allows me to display probe locations
on a
celfile image, but is it possible to go in the other direction i.e.
derive
the identities of probes related to a particular area on an image
(perhaps
specified interactively by pointing and clicking).
Thanks
--
----------------------------------------------------------------------
Matthew Hobbs
Garvan Institute of Medical Research
384 Victoria St Ph : (02) 9295 8327
Darlinghurst
http://www.garvan.org.au email: m.hobbs@garvan.org.au
It is possible (and hopefully not too complicated to use).
To get info about what is at x=100, y=50 you can do:
If you have the .CDF file around:
cdf <- read.cdffile("corresponding_cdf.cdf")
cdf@name.levels[cdf@name[100, 50]]
If you do not have it, one strategy is to get close to what
is above:
m.name <- matrix(as.integer(NA), nr=abatch@nrow, nc=abatch@ncol)
g.i <- indexProbes(abatch, which="both")
for (i in seq(along=g.i)) {
xy <- indices2xy(g.i[[i]], abatch=abatch)
m.name[xy] <- i
}
m.name.levels <- names(g.i)
m.name.levels[m.name[100, 50]]
For interactivity, you can implement easily something using the
R function 'locator'.
I am working on a more general scheme to shuttle from indices to
info(*),
and to positions to info... but I am not sure how fast this will come
true...
L.
(*: not only probe set id, but anything else one would fashion...)
On Wed, Jun 25, 2003 at 01:12:57PM +1000, Matthew Hobbs wrote:
> In affy the plotLocation function allows me to display probe
locations on a
> celfile image, but is it possible to go in the other direction i.e.
derive
> the identities of probes related to a particular area on an image
(perhaps
> specified interactively by pointing and clicking).
>
> Thanks
>
> --
>
----------------------------------------------------------------------
> Matthew Hobbs
>
> Garvan Institute of Medical Research
> 384 Victoria St Ph : (02) 9295 8327
> Darlinghurst
> http://www.garvan.org.au email: m.hobbs@garvan.org.au
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor@stat.math.ethz.ch
> https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor
I apologise for my question on com.braju.sma. I will write to Henrik.
Thanks again,
David
> Dear David,
>
> com.braju.sma is Henrik Bengtsson's very interesting package.
Although some
> of us work with Henrik on microarray problems, the com.braju.sma
package is
> not associated with bioconductor and isn't designed to work with
> bioconductor functions. You need to send any questions you have to
Henrik
> rather than to us!
>
> Best wishes
> Gordon
>
