[BioC] help with metaArray package

[Debashis Ghosh]

Dear Olga,

You are fitting a model to the data in which the poe.em$data should 
give you the  transformed data that are combinable across studies.  By 
definition, the entries in poe.em$data are not identical to the 
original data.

Debashis
-- 
Debashis Ghosh
Associate Professor of Biostatistics
University of Michigan
1420 Washington Heights
Ann Arbor, MI 48109-2029

email: ghoshd at umich.edu
phone: (734) 615-9824
fax: (734) 763-2215


Quoting "Olga K. Kamneva" <kamushkina at gmail.com>:

> Hello, all,
>
> I have some trouble with metaArray package.
> I'd like use this package to integrate data from different studies. But it
> give unexpected result.
> I have data set with 4 time points and 2 repeats for each time. I have an
> expression matrix after rma() prepossessing
>
>
>> str(b.rma)
> num [1:8641, 1:8] 4.95 3.95 5.48 7.74 5.29 ...
> - attr(*, "dimnames")=List of 2
>  ..$ : chr [1:8641] "Mm.4517" "Mm.226037" "Mm.338175" "Mm.362588" ...
>  ..$ : chr [1:8] "X00hr.r..CEL" "X00hr.CEL" "X08hr.r..CEL" "X08hr.CEL" ...
>>  b.rma[1:5,1:4]
>          X00hr.r..CEL X00hr.CEL X08hr.r..CEL X08hr.CEL
> Mm.4517       4.949947  4.899200     4.814221  4.839363
> Mm.226037     3.946517  4.018580     4.047994  4.000673
> Mm.338175     5.484870  5.458543     5.315563  5.375122
> Mm.362588     7.744338  7.761616     7.414228  7.394226
> Mm.1571       5.294042  5.108926     4.720241  4.825122
>>  clb <- c(1, 1, 0, 0, 0, 0, 0, 0)
>> b.poe.em<- poe.em(b.rma, cl = clb,  threshold=1e-05, every = 3000)
> Number of Samples: 8
> Number of Genes: 8641
> This model assumes that the samples are centered and scaled.
> 3000 genes fitted
> 6000 genes fitted
>> b.poe.em$data[1:5,1:4]
>           X00hr.r..CEL     X00hr.CEL X08hr.r..CEL    X08hr.CEL
> Mm.4517    0.0639460455  0.0096282914 -0.015324771 -0.003167604
> Mm.226037 -0.0221766586 -0.0003496772  0.001001762 -0.002457393
> Mm.338175 -0.0005617508 -0.0057849273 -0.211710552 -0.072901889
> Mm.362588  0.0153744690  0.0215050722 -0.097810670 -0.121062369
> Mm.1571    0.9081874141  0.9081608456  0.015922584  0.247076490
>>
>
>
> I have no idea why data after poe-transformation isn't conformed with
> original data.
> Looking forward any suggestions!
> Olga
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: 
> http://news.gmane.org/gmane.science.biology.informatics.conductor
>
>
>