[BioC] vsn on vsn

[Sean Davis]

On Fri, Sep 16, 2011 at 12:54 PM, Dr Balazs Gyorffy <zsalab2 at yahoo.com> wrote:
> Hi Sean,
> we have an online computational site www.kmplot.com. In the site we have
> integrated clinical information and several Affymetrix gene expression
> datasets from GEO, and the user is capable to perform a meta-analysis on
> these for predicting survival in breast and ovarian cancer.
> I would like to set up a similar system for miRNA analysis. However, for
> miRNA, only a very limited number of publications are available and these
> are on different platforms. For example prostate cancer looks very promising
> with two datasets with known relapse free survival: GSE26367 and GSE21032,
> these are on  Agilent and Illumina.
>
> We are also working in the lab with miRNA, but again, on a different
> platform (Invitrogen, see:
> http://www.ncbi.nlm.nih.gov/pubmed?term=gyorffy%20mirna).
>
> I see that the different platforms have the same miRNA oligos (let-7, etc).
> So the question is how to normalize them. One option is to use the raw data,
> but since the arrays are different, the processing will be different anyway.
> Another option is to take the normalized datasets from GEO and use a
> cross-project normalization. Earlier for mRNA (not miRNA) this worked fine
> for me using VSN. However, I see there that one of the datasets (the
> GSE21032) was already VSN normalized.

Hi, Balazs.

My experience has been that cross-project normalization, while
numerically possible, does not typically remove all the biases that
exist between different platforms, I am not sure what to tell you.  In
some cases (apparently the case for the data on your website), the
biases that likely remain are not too severe.  Sorry I couldn't be
more helpful.

Sean


> Best:
> Balazs
>
>
> -------------------------
> Balázs GYÖRFFY MD, PhD
> Research Laboratory for Pediatrics and Nephrology, Hungarian Academy of
> Sciences - Semmelweis University 1st Dept. of Pediatrics
> Bokay u. 53-54., Budapest H-1083
> Office: +36 1 459 1500 /52772
> Fax: +36 1 3036 077
> Email: zsalab2 at yahoo.com
> ________________________________
> Von: Sean Davis <sdavis2 at mail.nih.gov>
> An: Dr Balazs Gyorffy <zsalab2 at yahoo.com>
> Cc: bioconductor at r-project.org
> Gesendet: 14:35 Freitag, 16.September 2011
> Betreff: Re: [BioC] vsn on vsn
>
> On Fri, Sep 16, 2011 at 8:07 AM, Dr Balazs Gyorffy <zsalab2 at yahoo.com>
> wrote:
>> The problem is that one of them is Illumina and the other one is Agilent,
>> so
>> merging them at the raw data stage is very "hard".
>
> You should probably look at using a meta-analysis method rather than
> trying to merge "the numbers".  This type of cross-dataset analysis
> can be quite challenging and depends on the details of the two
> datasets and the questions one is trying to answer.  Defining how to
> do such an analysis in email without a lot more detail is not really
> possible.
>
> Sean
>
>
>> Best:
>> Balazs
>>
>> -------------------------
>> Balázs GYÖRFFY MD, PhD
>> Research Laboratory for Pediatrics and Nephrology, Hungarian Academy of
>> Sciences - Semmelweis University 1st Dept. of Pediatrics
>> Bokay u. 53-54., Budapest H-1083
>> Office: +36 1 459 1500 /52772
>> Fax: +36 1 3036 077
>> Email: zsalab2 at yahoo.com
>> ________________________________
>> Von: Sean Davis <sdavis2 at mail.nih.gov>
>> An: Dr Balazs Gyorffy <zsalab2 at yahoo.com>
>> Cc: "bioconductor at r-project.org" <bioconductor at r-project.org>
>> Gesendet: 13:39 Freitag, 16.September 2011
>> Betreff: Re: [BioC] vsn on vsn
>>
>> Hi, Balazs.
>>
>> You should probably merge at the "raw data" stage and then normalize
>> the data together as one batch.
>>
>> Sean
>>
>> On Fri, Sep 16, 2011 at 7:29 AM, Dr Balazs Gyorffy <zsalab2 at yahoo.com>
>> wrote:
>>> Hi All!
>>>
>>>
>>> I  want to merge the two different microarray datasets, one was VSN
>>> normalized and the other one quantile-normalized. Can I use a second VSN
>>> normalization? Any other normalization / related literature?
>>>
>>>
>>> Thanks:
>>>
>>> Balazs
>>>
>>>
>>> -------------------------
>>> Balázs GYÖRFFY MD, PhD
>>> Research Laboratory for Pediatrics and Nephrology, Hungarian Academy of
>>> Sciences - Semmelweis University 1st Dept. of Pediatrics
>>> Bokay u. 53-54., Budapest H-1083
>>> Office: +36 1 459 1500 /52772
>>> Fax: +36 1 3036 077
>>> Email: zsalab2 at yahoo.com
>>>
>>>        [[alternative HTML version deleted]]
>>>
>>>
>>> _______________________________________________
>>> Bioconductor mailing list
>>> Bioconductor at r-project.org
>>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>>> Search the archives:
>>> http://news.gmane.org/gmane.science.biology.informatics.conductor
>>>
>>
>>
>>
>
>
>