[BioC] how to use normalizeForPrintorder accounting for duplicates when there are spots that are not duplicated in the array?

[Gordon K Smyth]

Dear Artur,

Any settings for layout that lead to same values for printorder(layout)$printorder will give the
same results in normalizeForPrintorder().

Best wishes
Gordon

On Thu, June 21, 2007 11:17 pm, Artur Veloso wrote:
> Dear Gordon,
>
> thank you very much for your reply. I think my mistake in those regards was
> caused by the fact that the function printorder allows one to specify the
> number of duplicates in the array.
>
> In an unrelated issue, I've been using limma to analyze single channel cDNA
> arrays and many times ran into errors message like "List must contain
> components R and G" generated by the lack of the second channel on my
> arrays. My solution was to change the code for almost all the functions that
> I used on the pre-processing stage. Basically, I just removed the parts of
> the code that referred to the non-existing channel.
> I just wanted to bring awareness to this issue and offer to e-mail you the
> changes I made on the functions to make them more suitable for my situation.
>
> All the best,
>
> Artur Veloso
>
> On 6/21/07, Gordon K Smyth <smyth at wehi.edu.au> wrote:
>>
>> Dear Artur,
>>
>> You question doesn't really make sense, because print-order normalization
>> has nothing to do with
>> duplicated spots.  The identities of the probes are ignored in such a
>> normalization.  In any case,
>> you can't move spots around and expect the function to still keep track
>> correctly of the print
>> order.  In the limma package, handling of duplicated spots is done
>> post-normalization.
>>
>> Best wishes
>> Gordeon
>>
>> > Date: Wed, 20 Jun 2007 15:40:18 -0400
>> > From: "Artur Veloso" <abveloso at gmail.com>
>> > Subject: [BioC] how to use normalizeForPrintorder accounting for
>> >       duplicates      when there are spots that are not duplicated in
>> the array?
>> > To: "Bioconductor List" <bioconductor at stat.math.ethz.ch>
>> >
>> > Hi all,
>> >
>> > I am very new to microarray analysis and I'm running into a few issues
>> with
>> > normalizations for the print order. The arrays I am working with are
>> > "home-made" and the spots are duplicated and separated by columns.
>> >
>> > My problem arises from the fact that there are non-informative spots (
>> i.e.
>> > landing lights and empty spots) that are not duplicated. Would it be
>> > acceptable for me to take all these uninformative spots and shift them
>> to
>> > the end of my spots list so that they didn't interfere with the
>> informative
>> > and properly duplicated spots? Could this interfere with the results
>> that
>> > I'd get from the function normalizeForPrintorder?
>> > If shifting the position of my spots like that isn't a good idea, does
>> > anyone have a suggestion to me?
>> >
>> > Thank you all very much,
>> >
>> > Artur Veloso