[BioC] SpotTypes doesn't match

Gordon Smyth smyth at wehi.edu.au
Sun Sep 11 09:37:51 CEST 2005


If you type

   names(RGgpr$genes)

you will see that there is no column called 'controltype' in the annotation 
component your data object. The problem is that 'controltype' is not a 
standard GenePix annotation column. Hence limma will not know to read it in 
unless you tell it to by including this column in the 'annotation' argument 
of read.maimages(). That is, you need to use

   RGgpr <- read.maimages(...,  annotation=c("Name","controltype"))

Gordon

>To: kevin Lin <khlin at odin.mdacc.tmc.edu>
>From: Gordon Smyth <smyth at wehi.edu.au>
>Subject: [BioC] SpotTypes doesn't match
>Cc: bioconductor at stat.math.ethz.ch
>Date: Thu, 08 Sep 2005 10:30:21 +1000
>
>I suspect you will find that there is no column called "controltype" in 
>your GPR files. Perhaps check the capitalisation.
>
>Gordon
>
>>Date: Tue, 6 Sep 2005 14:47:22 -0500
>>From: kevin Lin <khlin at odin.mdacc.tmc.edu>
>>Subject: [BioC] SpotTypes doesn't match
>>To: <Bioconductor at stat.math.ethz.ch>
>>Message-ID: <a06002001bf439982ab56@[143.111.133.119]>
>>Content-Type: text/plain; charset="us-ascii" ; format="flowed"
>>
>>Hi BioC,
>>
>>I am using Limma, the problem occurs in my 2nd object RGgpr$genes$Status.
>>
>> >  SpotTypes
>>    SpotType controltype Name  color
>>1  cDNA           *        *         black
>>2  genes       FALSE    *         grey
>>3  positive    positive    *         blue
>>4  ignore      ignore      *         orange
>>5 notprobe    notprobe  0         yellow
>>
>> >  RGgpr$genes$Status <- controlStatus(SpotTypes, RGgpr$genes)
>>Matching patterns for: Name
>>Found 11592 cDNA
>>Found 11592 genes
>>Found 11592 positive
>>Found 11592 ignore
>>Found 664 notprobe
>>Setting attributes: values controltype color
>>
>>I cut and pasted two columns (Name and controltype) of *.gpr files to
>>confirm the different numbers of controltypes.
>> >table(read.table("twoColumns.txt",sep='\t',header=F))
>>FALSE   ignore notprobe positive
>>    8564     2112      664      252
>>
>>Q1, Why did they give me the same # 11592 ? Is it because I used wild
>>card "*" ? Can this be fixed?
>>Q2, If so, what can i do with the SpotType.txt to fixed the problem ?
>>
>>Thanks for any help
>>
>>Kevin



More information about the Bioconductor mailing list