qualityParameters {arrayMagic} | R Documentation |
arrayDataObject |
object of type arrayData ; default: missing |
exprSetArrayObject |
object of type exprSetArray ; default: missing |
spotIdentifier |
character string, column name used in getSpotAttr(arrayDataObject); the column is used to determine spot replicas; default: "Name" |
hybNameColumn |
character string, column name used in getHybAttr(arrayDataObject); the column is used to extract the names of the hybridisations; if not available the order is taken instead; default: "slideNumber" |
identifiersToBeSkipped |
vector of character strings of spot identifiers to be excluded from calculations |
resultFileName |
character string; results are stored in a file if supplied; default: missing |
verbose |
logical; default TRUE |
returns a list of results, i.e. a data frame qualityParameters
containing
several scores for each hybridisation, a matrix slideDistance
and
an integer replicateSpots
, i.e. the number of spot replicates.
The matrix slideDistance
contains a distance measure (similarity) for each pair
of slides$_{ij}$, i.e. the median absolute deviation taken over all spots of the
difference of the "log-ratios" ("log-ratios": the difference of the expression values
in the two channels on the slide).
A brief summary of all parameters given the data frame qualityParameters
:
width
a robust measure of the variance, i.e. the median absolute deviation
of the difference of the normalised channels taken over all spots
medianDistance
a robust measure for the typical distance (similarity) of one slide
with all other slides, i.e. the median of the "distances"
between slides (c.f. slideDistance
))
correlation
of the expression values between the two channels of the slide taken over all spots
meanIntensityCy3
the mean taken over all spots of raw data channel Cy3
meanIntensityCy5
the mean taken over all spots of raw data channel Cy5
meanIntensity
mean taken over all spots of raw data of both channels,
spotSimilarityCy3
the standard deviation of all spot replicates for each unique identifier
of the normalised channel Cy3 is calculated and the
median standard deviation
is taken as spotSimilarityCy3; if duplicate spots have been detected, i.e.
replicateSpots == 2
, the correlation is calculated instead
spotSimilarityCy3
the standard deviation of all spot replicates for each unique identifier
of the normalised channel Cy3 is calculated and the
median standard deviation
is taken as spotSimilarityCy3; if duplicate spots have been detected, i.e.
replicateSpots == 2
, the correlation is calculated instead
cy3vsAllCy3
and cy5vsAllCy5
the correlation between each channel is measured against the
averaged (median) channel over all hybridisations
(e.g. a virtual reference)
Andreas Buness <a.buness@dkfz.de>
spotIdentifierVec <- c("A","A","Blank","B","B","Blank") hybNames <- "H1" R1 <- N1 <- c(1,1,9,2,2,10) R2 <- N2 <- c(2,2,7,4,4,8) rawDataIntensities <- array(0, dim=c(6,2,1)) rawDataIntensities[,1,] <- R1 rawDataIntensities[,2,] <- R2 dimnames(rawDataIntensities) <- list(NULL, c("Cy3","Cy5"), NULL) spotattr=data.frame(Name=I(spotIdentifierVec)) hybattr=data.frame(slideNumber=I(hybNames)) arrayDataObject <- new("arrayData", intensities=rawDataIntensities, hybattr=hybattr, spotattr=spotattr) indCy3=1 indCy5=2 channels <- matrix( c(indCy3,indCy5), nrow=length(indCy3), byrow=FALSE ) colnames(channels) <- c("Cy3","Cy5") exprSetArrayObject <- new("exprSetArray", exprs=matrix(c(R1,R2), nrow=6, byrow=FALSE), phenoData= new("phenoData", pData=data.frame(matrix(0,nrow=2,ncol=1)), varLabels=list(rep("varLabel1",1))), channels=channels) Re1 <- qualityParameters(arrayDataObject=arrayDataObject, exprSetArrayObject=exprSetArrayObject, identifiersToBeSkipped= "Blank") stopifnot(all.equal.numeric(as.numeric(Re1$qualityParameters["H1",c("correlation")]),c(1))) stopifnot(Re1$replicateSpots==2)