[BioC] Reading Agilent files using read.maimages

[Steve Taylor]

Hi Sean,

>>A typical header for one of these raw files is
>>
>>CompositeSequence Identifier    Database ebi.ac.uk:Database:embl        Database ebi.ac.uk:Database:ensembl     Database ebi.ac.uk:Database:locus       Database ebi.ac.uk:Database:refseq      Database 
>>ebi.ac.uk:Database:tigr_thc    Database www.chem.agilent.com:Database:agc      Database www.chem.agilent.com:Database:agp      Feature coordinates: metaColumn metaRow column  row     Reporter control 
>>type   Reporter group  Reporter identifier     Reporter name   Reporter sequence type  H_C1 A_mock A and B/FEATURES    H_C1 A_mock A and B/FeatureNum  H_C1 A_mock A and B/gbpri       H_C1 A_mock A and 
>>B/gp  H_C1 A_mock A and B/sp  H_C1 A_mock A and B/ProbeUID    H_C1 A_mock A and B/ControlType H_C1 A_mock A and B/ProbeName   H_C1 A_mock A and B/GeneName    H_C1 A_mock A and B/SystematicName 
>>H_C1 A_mock A and B/Description H_C1 A_mock A and B/LogRatio    H_C1 A_mock A and B/LogRatioError       H_C1 A_mock A and B/PValueLogRatio      H_C1 A_mock A and B/gSurrogateUsed      H_C1 A_mock A 
>>and B/rSurrogateUsed      H_C1 A_mock A and B/gIsFound    H_C1 A_mock A and B/rIsFound    H_C1 A_mock A and B/gProcessedSignal    H_C1 A_mock A and B/rProcessedSignal    H_C1 A_mock A and 
>>B/gProcessedSigError  H_C1 A_mock A and B/rProcessedSigError  H_C1 A_mock A and B/gNumPixOLHi H_C1 A_mock A and B/rNumPixOLHi H_C1 A_mock A and B/gNumPixOLLo H_C1 A_mock A and B/rNumPixOLLo H_C1 
>>A_mock A and B/gNumPix     H_C1 A_mock A and B/rNumPix     H_C1 A_mock A and B/gMeanSignal H_C1 A_mock A and B/rMeanSignal H_C1 A_mock A and B/gMedianSignal       H_C1 A_mock A and B/rMedianSignal 
>>    H_C1 A_mock A and B/gPixSDev    H_C1 A_mock A and B/rPixSDev    H_C1 A_mock A and B/gBGNumPix   H_C1 A_mock A and B/rBGNumPix   H_C1 A_mock A and B/gBGMeanSignal       H_C1 A_mock A and 
>>B/rBGMeanSignal       H_C1 A_mock A and B/gBGMedianSignal     H_C1 A_mock A and B/rBGMedianSignal     H_C1 A_mock A and B/gBGPixSDev  H_C1 A_mock A and B/rBGPixSDev  H_C1 A_mock A and B/gNumSatPix 
>>H_C1 A_mock A and B/rNumSatPix  H_C1 A_mock A and B/gIsSaturated        H_C1 A_mock A and B/rIsSaturated        H_C1 A_mock A and B/PixCorrelation      H_C1 A_mock A and B/BGPixCorrelation    H_C1 
>>A_mock A and B/gIsFeatNonUnifOL    H_C1 A_mock A and B/rIsFeatNonUnifOL    H_C1 A_mock A and B/gIsBGNonUnifOL      H_C1 A_mock A and B/rIsBGNonUnifOL      H_C1 A_mock A and B/gIsFeatPopnOL       H_C1 
>>A_mock A and B/rIsFeatPopnOL       H_C1 A_mock A and B/gIsBGPopnOL H_C1 A_mock A and B/rIsBGPopnOL H_C1 A_mock A and B/IsManualFlag        H_C1 A_mock A and B/gBGSubSignal        H_C1 A_mock A and 
>>B/rBGSubSignal        H_C1 A_mock A and B/gBGSubSigError      H_C1 A_mock A and B/rBGSubSigError      H_C1 A_mock A and B/BGSubSigCorrelation H_C1 A_mock A and B/gIsPosAndSignif     H_C1 A_mock A and 
>>B/rIsPosAndSignif     H_C1 A_mock A and B/gPValFeatEqBG       H_C1 A_mock A and B/rPValFeatEqBG       H_C1 A_mock A and B/gNumBGUsed  H_C1 A_mock A and B/rNumBGUsed  H_C1 A_mock A and B/gIsWellAboveBG 
>>      H_C1 A_mock A and B/rIsWellAboveBG      H_C1 A_mock A and B/IsUsedBGAdjust      H_C1 A_mock A and B/gBGUsed     H_C1 A_mock A and B/rBGUsed     H_C1 A_mock A and B/gBGSDUsed   H_C1 A_mock A and 
>>B/rBGSDUsed   H_C1 A_mock A and B/IsNormalization     H_C1 A_mock A and B/gDyeNormSignal      H_C1 A_mock A and B/rDyeNormSignal      H_C1 A_mock A and B/gDyeNormError       H_C1 A_mock A and 
>>B/rDyeNormError       H_C1 A_mock A and B/DyeNormCorrelation  H_C1 A_mock A and B/ErrorModel
> 
> 
> This is not an Agilent Raw Data file, I do not think.  The column names
> are similar, but ArrayExpress has significantly changed the file from
> its original format.  That said, the columns with "LogRatio",
> "rProcessedSignal" and "gProcessedSignal" are the columns of interest
> that have already been background corrected and, typically, a
> normalization method applied (not sure which one without some more
> description of the scanner settings).
> 

Ok. Thanks. That's useful information. In the protocols section of AE it says 'Default settings' 
(http://www.ebi.ac.uk/aerep/details?class=MAGE.Experiment_protocols&criteria=Experiment%3D921408317&contextClass=MAGE.Protocol&templateName=Protocol.vm). If that means it has been normalised I will 
have a look at LogRatio, rProcessedSignal and gProcessedSignal, though it would be nice to know how it had been processed...

> 
>>Does this look correct? How do I get access to the intensities, for example to do a boxplot?
> 
> 
> I'm not sure if the files loaded correctly, given my comments above.
> RG$R and RG$G contain the Red and Green intensities, if it loaded correctly.
> 

That's what I thought. Thanks for the advice,

Steve