James Anderson wrote:
> Hi,
>
> I am analyzing about 30 agilent two color arrays. One simple
question is that, in terms of the downstream analysis such as finding
differentially expressed genes, what are the features should I use
(such as log ratio, mean intensity of red and green, etc)? I have some
experience in Affy, but almost no experience in agilent.
>
> In addition, I am trying to use arrayMagic from Bioconductor to do
quality control and pre-processing of my agilent arrays. From the
example arrayMagic gave in the manual, it seems that the format of the
example data is somewhat different than agilent data format.
>
> I saw the following header file
>
> HEADER SPOT GRID TOP LEFT BOT
RIGHT ROW COL CH1I CH1B CH1AB CH2I
CH2B CH2AB SPIX BGPIX EDGE RAT2
MRAT REGR CORR LFRAT CH1GTB1 CH2GTB1
CH1GTB2 CH2GTB2 CH1EDGEA CH2EDGEA FLAG
CH1KSD CH1KSP CH2KSD CH2KSP
> While my agilent arrays has the following header information:
>
> FeatureNum Row Col Start Sequence
ProbeUID ControlType ProbeName GeneName
PositionX PositionY LogRatio LogRatioError
PValueLogRatio gSurrogateUsed rSurrogateUsed
gIsFound rIsFound gProcessedSignal rProcessedSignal
gProcessedSigError rProcessedSigError gNumPixOLHi
rNumPixOLHi gNumPixOLLo rNumPixOLLo gNumPix
rNumPix gMeanSignal rMeanSignal gMedianSignal
rMedianSignal gPixSDev rPixSDev gBGNumPix
rBGNumPix gBGMeanSignal rBGMeanSignal
gBGMedianSignal rBGMedianSignal gBGPixSDev
rBGPixSDev gNumSatPix rNumSatPix gIsSaturated
rIsSaturated PixCorrelation BGPixCorrelation
gIsFeatNonUnifOL rIsFeatNonUnifOL gIsBGNonUnifOL
rIsBGNonUnifOL gIsFeatPopnOL rIsFeatPopnOL
> gIsBGPopnOL rIsBGPopnOL IsManualFlag
gBGSubSignal rBGSubSignal gBGSubSigError
rBGSubSigError BGSubSigCorrelation gIsPosAndSignif
rIsPosAndSignif gPValFeatEqBG rPValFeatEqBG
gNumBGUsed rNumBGUsed gIsWellAboveBG rIsWellAboveBG
gBGUsed rBGUsed gBGSDUsed rBGSDUsed
IsNormalization gDyeNormSignal rDyeNormSignal
gDyeNormError rDyeNormError DyeNormCorrelation
ErrorModel xDev gSpatialDetrendIsInFilteredSet
rSpatialDetrendIsInFilteredSet gSpatialDetrendSurfaceValue
rSpatialDetrendSurfaceValue
>
> How should I use arrayMagic on my data?
>
I haven't used arrayMagic for Agilent data. You can use limma and
read.maimages() to load your data. As for what columns to use, the
logRatio column of course contains a ratio, while the columns
rProcessedSignal and gProcessedSignal probably contain the
background-corrected intensities. You will probably want to glance at
the Agilent technical manual about what the columns represent, in more
detail. As for what columns you will want to use in your analysis,
you
will need to be specific about your experimental design. If you used
a
common reference, then you can simply treat the logRatio in the same
way
as your affymetrix arrays for the purposes of analysis.
Sean
