[BioC] output from "getChipInfo" prevents Illumina ProbeID conversion

[Pan Du]

Hi Chris

The "summary(lumi_vst)" returns the information shown in the header of
your input data file. While the getChipInfo returns the best match to
the tables in the HumanIDMapping package. To save the storage space,
the HumanIDMapping package does not include all sub-version tables. If
two sub-versions have the same probes, it only keep the latest
version. So you can use "HumanHT12_V3_0_R3_11283641_A" to map your
data.

 If you provide your own annotationFile,  the function assumes your
data is annotated with Illumina ProbeID or TargetID.  Since your data
uses "Array_Address_Id", it reports error.



Pan

On Fri, Jun 8, 2012 at 2:11 PM, Chris Gaiteri <gaiteri at gmail.com> wrote:
> Hello Dr. Du,
>
> There is a difference between the output of "getChipInfo" and "summary"
> functions applied the same lumi object, which is preventing us from mapping
> the Illumina ProbeID's to gene symbols or other useful output.  Have you
> encountered this situation before?  I do appreciate the time you put into
> these packages.  Annotated output below.
>
> Regards,
> Chris Gaiteri
>
> "lumi_vst" is the Beadstudio output, processed through the lumi package.
>> summary(lumi_vst)
>
> Summary of data information:
>          Data File Information:
>                 Illumina Inc. GenomeStudio version 1.0.6
>                 Normalization = none
>                 Array Content = HumanHT-12_V3_0_R2_11283641_A.bgx.xml
>
> (the above does not match the output of "getChipInfo" for instance "R2" vs
> "R3" and the extra chip version detected by getChipInfo)
>
>
>
>> getChipInfo(lumi_vst)
> $chipVersion
> [1] "HumanHT12_V3_0_R3_11283641_A" "HumanWG6_V3_0_R3_11282955_A"
>
> $species
> [1] "Human"
>
> $IDType
> [1] "Array_Address_Id" "Array_Address_Id"
>
> $chipProbeNumber
> [1] 48803
>
> $inputProbeNumber
> [1] 48803
>
> $matchedProbeNumber
> [1] 48803
>
>
> (attempting to generate nuID's fails, using what I know is the correct
> annotation file)
> lumi_vst = addNuID2lumi(lumi_vst,
> annotationFile='HumanHT-12_V3_0_R2_11283641_A.txt')
> Error in addNuID2lumi(lumi_vst, annotationFile =
> "HumanHT-12_V3_0_R2_11283641_A.txt") : The annotation file does not match
> the data!
>
> (In fact "addNuID2lumi" does not consider than any annotation file matches -
> no matter which suggested version of annotation I use)
>
> other attempts to convert to useful ID's also fail:
>> featureNames(lumi_vst)[1:10]
>  [1] "6450255" "2570615" "6370619" "2600039" "2650615" "5340672" "1090041"
> "6380561" "7570255" "4920477"
>> nuIDs = IlluminaID2nuID(featureNames(lumi_vst))
> Warning message:
> In if (!is.na(chipInfo$IDType)) { :
>   the condition has length > 1 and only the first element will be used
>
> (I think this comes from the two chip types that are listed under
> getChipInfo$chipVersion )
>
>
>> nuIDs[1:10]
>  [1] "ILMN_5579"   "ILMN_175569" "ILMN_18893"  "ILMN_18532"  "ILMN_6386"
> "ILMN_6386"   "ILMN_2229"   "ILMN_2229"
>  [9] "ILMN_19974"  "ILMN_31305"
>
> (these to not appear to be NuIDs and prevent further processing with
> lumiHumanIDMapping package)