After checking your data, I found the data has already log2
VST transformation should be on the raw scale. So please convert your
to raw scale first, then you can apply vst.
exprs(x.lumi) <- 2^(exprs(x.lumi))
I will also add checking code in the lumi package to give warning
the cases like yours. Tell me if the problem still exists.
On 2/27/11 8:04 AM, "Sarah Allen" <sa413 at="" cam.ac.uk=""> wrote:
> Hello, thanks for your reply. I have attached the session info and
> file of lumiB like you said. Let me know if you need anything else.
> Thank you very much for your time. Sarah
> On 26 Feb 2011, at 15:35, Pan Du wrote:
>> Hi Sarah
>> The information you provided is too limited, I cannot answer your
>> At least, you need to provide sessionInfo() and the information of
>> You can also send me a subset of lumi.B, for example,
>> test = lumi.B[1:5000,]
>> save(test, file='test.Rdata')
>> and send me 'test.Rdata' file.
>> On 2/26/11 5:00 AM, "bioconductor-request at r-project.org"
>> <bioconductor-request at="" r-project.org=""> wrote:
>>> Date: Fri, 25 Feb 2011 11:39:12 +0000
>>> From: Sarah Allen <sa413 at="" cam.ac.uk="">
>>> To: bioconductor at r-project.org
>>> Subject: [BioC] Lumi: LumiT vst transform error
>>> Message-ID: <0C717235-749C-4C91-85F3-9F19AED64DBA at cam.ac.uk>
>>> Content-Type: text/plain; charset=US-ASCII; format=flowed;
>>> Hello, I am having a problem performing vst normalisation on my
>>> I get the following error message that I can't make head or tail
>>>> lumi.T <- lumiT(lumi.B)
>>> Perform vst transformation ...
>>> 2011-02-25 11:07:29 , processing array 1
>>> Error in lm.fit(x, y, offset = offset, singular.ok =
>>> singular.ok, ...) :
>>> 0 (non-NA) cases
>>> I have searched the archives and found that people have asked the
>>> question before and the advice has been to update their
>>> package. All my packages are up to date, I have made sure that all
>>> values are positive using lumiB "forcePositive", my boxplots all
>>> fine (there is nothing intrinsically weird about my data) ... I am
>>> Can anyone help?
>>> Many thanks, Sarah Allen