Hi All,
I am trying to perform loess normalization on spotted oligo chip data
which have only one channel, instead of two-channel in cDNA data. I
would also like to use the print-tip normalization option. I
understand
I could use the loess normalization function from the "affy" package,
but I don't believe it takes into account the print-tip. Does anybody
has experience in this issue?
thank you
Tzu
> Hi All,
>
> I am trying to perform loess normalization on spotted oligo chip
data
> which have only one channel, instead of two-channel in cDNA data. I
> would also like to use the print-tip normalization option. I
understand
> I could use the loess normalization function from the "affy"
package,
> but I don't believe it takes into account the print-tip. Does
anybody
> has experience in this issue?
See the normalizeBetweenArrays() function in the limma package for
various
methods. Print-tip loess doesn't make nearly as much sense for single
channel as for two colour data. We use quantile normalization as the
default for single-channel data.
Gordon
We are very interested in eliminating the print-tip effects for our
single channel data. Do you know if any group have tried to
accomplish
that on the single -channel oligo chips data?
Thank you
Tzu
Gordon K Smyth wrote:
>>Hi All,
>>
>>I am trying to perform loess normalization on spotted oligo chip
data
>>which have only one channel, instead of two-channel in cDNA data. I
>>would also like to use the print-tip normalization option. I
understand
>> I could use the loess normalization function from the "affy"
package,
>>but I don't believe it takes into account the print-tip. Does
anybody
>>has experience in this issue?
>>
>>
>
>See the normalizeBetweenArrays() function in the limma package for
various
>methods. Print-tip loess doesn't make nearly as much sense for
single
>channel as for two colour data. We use quantile normalization as the
>default for single-channel data.
>
>Gordon
>
>
>
>
>
>
--
Tzu L. Phang, Ph.D.
303-315-1583
http://compbio.uchsc.edu/Hunter_lab/Phang
[[alternative HTML version deleted]]
Hi Tzulip,
you can try the function vsn in the package of the same name. If y is
a
matrix of size (no.probes) x (no.chips), and pins is an integer vector
of length (no.probes) with values 1, 2, .... that code for the print-
tip
ID, then
ny1 <- vsn(y, strata=pins)
should do the job. You can compare this to
ny2 <- vsn(y)
for a single-channel normalization "without" print-tips. For details,
please see the vignette. I'd be interested in your experience with
this.
library(vsn)
openVignette("vsn")
Best wishes
Wolfgang
Tzulip Phang wrote:
> We are very interested in eliminating the print-tip effects for our
> single channel data. Do you know if any group have tried to
accomplish
> that on the single -channel oligo chips data?
>
> Thank you
>
> Tzu
>
> Gordon K Smyth wrote:
>
-------------------------------------
Wolfgang Huber
Division of Molecular Genome Analysis
German Cancer Research Center
Heidelberg, Germany
Phone: +49 6221 424709
Fax: +49 6221 42524709
Http: www.dkfz.de/abt0840/whuber
