[BioC] #Identify differentially expressed genes

Pan Du dupan.mail at gmail.com
Tue Aug 28 18:02:10 CEST 2012


>
> I was wondering on the scale of the heatmap, usually values ranges
> from -3 to 3, but data VST transformed in my example ranges from 7 to
> 14, with a median value of 7.5...
> is it correct to plot a heatmap with VST transfmed data? or is
> preferable to perfrom variance stabilization using log2 transform?
> If is correct to use vst transformed data how do I refer to the scale?

The scale of heatmap in the range between -3 and 3 is because by
default "heatmap" function will scale each row to 0 mean and 1
standard deviation. The purpose of this is to visualize the relative
changes of the patterns. But in the meantime, the original scale
information is lost. So the heatmap scale has no direct relation with
vst or log2 transform.

Pan

>
>
>
>
> 2012/8/23 Pan Du <dupan.mail at gmail.com>:
>> Hi Paolo
>>
>> It's pretty normal that some probes cannot map to any genes because
>> Human genome annotation keeps updating but the probe design is
>> unchanged. So usually you can ignore those NA probes. If you are
>> really interested in them, you can easily convert the nuID to probe
>> sequence (use id2seq function) and map them to genome or refseq.
>> I will update the vignette later to avoid such confusion.
>>
>> Pan
>>
>> On Thu, Aug 23, 2012 at 3:26 AM, Paolo Kunderfranco
>> <paolo.kunderfranco at gmail.com> wrote:
>>> Dear All,
>>> I am working with lumi / limma package to detect differentially expressed
>>> genes between two or more samples.
>>> I was wondering why when I add geneSymbol and geneName to my Illumina
>>> probelist most of them (around 10%)are not called and remained NA, for
>>> instance (last row):
>>>
>>>
>>> if (require(lumiMouseAll.db) & require(annotate)) {
>>>                geneSymbol <- getSYMBOL(probeList, 'lumiMouseAll.db')
>>>                geneName <- sapply(lookUp(probeList, 'lumiMouseAll.db',
>>> 'GENENAME'), function(x) x[1])
>>>                fit1_2$genes <- data.frame(ID= probeList,
>>> geneSymbol=geneSymbol, geneName=geneName, stringsAsFactors=FALSE)
>>>           }
>>>
>>>
>>> 7671  69fpKOOuFduFbAjNVU     Dppa5a    developmental pluripotency
>>> associated 5A  7.828381  9.333743 149.31710 2.773571e-18 6.144846e-14
>>> 29.80858
>>> 16014 QpWgiAmByT4gW7iui0     Pou5f1 POU domain, class 5, transcription
>>> factor 1  5.305532  8.633706 103.85793 1.098423e-16 8.143726e-13
>>> 27.72832
>>> 20450 HlUzpCHheswfSZNdQo        Trh               thyrotropin
>>> releasing hormone  5.603441  8.761965 103.81774 1.102739e-16
>>> 8.143726e-13 27.72571
>>> 7670  o7Ah_nzF7JdZOTtd9U      Dppa4     developmental pluripotency
>>> associated 4  5.300619  8.626239  99.82457 1.640729e-16 9.087587e-13
>>> 27.45790
>>> 7672  xjn0tTp4isUXmUkAKI     Dppa5a    developmental pluripotency
>>> associated 5A  7.663922  9.439668  97.09091 2.173661e-16 9.631491e-13
>>> 27.26346
>>> 17719 ZXvxHuC6s3xogRFJfo      Sall4                     sal-like 4
>>> (Drosophila)  4.456642  8.585243  90.39110 4.484584e-16 1.655932e-12
>>> 26.74469
>>> 14084 06jqfFxe5_X97NRXuk       Myl3                 myosin, light
>>> polypeptide 3 -7.736059 13.128014 -88.39591 5.622067e-16 1.779384e-12
>>> 26.57755
>>> 8757  oii7mSFyrr_AMWODH0       <NA>
>>>     <NA>  4.608167  8.438770  78.65631 1.833459e-15 5.077535e-12
>>> 25.66512
>>>
>>> any ideas?
>>> thanks
>>> paolo
>>>



More information about the Bioconductor mailing list