SpotData: The SpotData class

Description Usage Arguments Fields and Methods Details About IQR More on background estimates Author(s) References Examples

Description

Package: aroma
Class SpotData

Object
~~|
~~+--MicroarrayData
~~~~~~~|
~~~~~~~+--SpotData

Directly known subclasses:

public static class SpotData
extends MicroarrayData

Creates an SpotData object. If the data frame data is empty or NULL, the object will be empty.

Usage

1

Arguments

layout

A Layout object specifying the spot layout of the slides in this data set.

Fields and Methods

Methods:

append -
as.RawData -
extractLayout -
getArea -
getArrayAspectRatio -
getArrayBottom -
getArrayHeight -
getArrayLeft -
getArrayRight -
getArrayTop -
getArrayWidth -
getBackground -
getBgArea -
getCircularity -
getForeground -
getForegroundSD Gets (an approximation of) the standard deviation of the foreground pixels.
getForegroundSE Gets the standard error of the foreground pixels.
getPerimeter -
getRawData Gets the raw intensites from the SpotData structure.
getSNR -
getSpotPosition Gets physical positions of the spots.
log -
normalizeGenewise -
plotSpatial Creates a spatial plot of a slide.
plotSpatial3d -
read Reads several Spot files into a SpotData object.
write Write a SpotData object to file.

Methods inherited from MicroarrayData:
addFlag, append, applyGenewise, applyGroupwise, applyPlatewise, applyPrintdipwise, applyPrinttipwise, as.character, as.data.frame, boxplot, clearCache, clearFlag, createColors, dataFrameToList, equals, extract, getBlank, getCache, getChannelNames, getColors, getExcludedSpots, getExtra, getExtreme, getFieldNames, getFlag, getInclude, getLabel, getLayout, getProbeWeights, getSignalWeights, getSlideNames, getSlidePairs, getSpotPosition, getSpotValue, getTreatments, getView, getWeights, getWeightsAsString, hasExcludedSpots, hasLayout, hasProbeWeights, hasSignalWeights, hasWeights, highlight, hist, isFieldColorable, keepSlides, keepSpots, listFlags, lowessCurve, nbrOfDataPoints, nbrOfFields, nbrOfSlides, nbrOfSpots, nbrOfTreatments, normalizePlatewise, normalizePrintorder, normalizeQuantile, plot, plotDensity, plotGene, plotPrintorder, plotReplicates, plotSpatial, plotSpatial3d, plotXY, points, putGene, putSlide, qqnorm, quantile, range, range2, read, readHeader, readToList, removeSlides, removeSpots, resetProbeWeights, resetSignalWeights, select, seq, setCache, setExcludedSpots, setExtra, setFlag, setLabel, setLayout, setProbeWeights, setSignalWeights, setSlideNames, setTreatments, setView, setWeights, size, str, subplots, summary, text, updateHeader, validateArgumentChannel, validateArgumentChannels, validateArgumentGroupBy, validateArgumentSlide, validateArgumentSlides, validateArgumentSpotIndex, validateArgumentWeights, write, writeHeader

Methods inherited from Object:
$, $<-, [[, [[<-, as.character, attach, attachLocally, clearCache, clone, detach, equals, extend, finalize, gc, getEnvironment, getFields, getInstanciationTime, getStaticInstance, hasField, hashCode, ll, load, objectSize, print, save

Details

A Spot file contains spot information for each spot on a single microarray slide. It consists of a header followed by a unspecified number of rows. The header contains 1+30 labels, and each row contains 31 fields. Each row corresponds to one spot. The fields are:

\item

<NO NAME>row number

\item

indexsspot number on slide. Range [0,N] in N.

\item

grid.rgrid row number. Range [1,GR] in Z+. \itemgrid.cgrid column number. Range [1,GC] in Z+. \itemspot.rspot row (within grid) number. Range [1,SR] in Z+. \itemspot.cspot column (within grid) number. Range [1,SC] in Z+.

\item

areathe number of foreground pixels. Range [0,MAXAREA] in N

\item

Gmeanthe average of the foreground pixel values. Range [0,65535] in R \itemGmedianthe median of the foreground pixel values. Range [0,65535] in N \itemGIQRthe inter quartile range (a robust measure of spread) of the logged foregroud pixel values. Range [0,16]+Inf in R,NA \itemRmeanthe average of the foreground pixel values. Range [0,65535] in R \itemRmedianthe median of the foreground pixel values. Range [0,65535] in N \itemRIQRthe inter quartile range (a robust measure of spread) of the logged foregroud pixel values. Range [0,16]+Inf in R,NA

\item

bgGmeanmean green background intesity. Range [0,65535] in R \itembgGmedmedian green background intesity. Range [0,65535] in N \itembgGSDstandard deviation for the green background. Range [0,65535]+Inf in R \itembgRmeanmean red background intesity. Range [0,65535] in R \itembgRmedmedian red background intesity. Range [0,65535] in N \itembgRSDstandard deviation for the red background. Range [0,65535]+Inf in R

\item

valleyGthe background intesity estimate from the local background valley method S.valley. Range [0,65535] in N \itemvalleyRthe background intesity estimate from the local background valley method S.valley. Range [0,65535] in N

\item

morphGgreen background estimate using morphological opening (erosion-dilation). Range [0,65535] in N \itemmorphG.erodegreen background estimate using morphological erosion. Range [0,65535] in N \itemmorphG.close.opengreen background estimate using morphological closing-opening (dilation-erosion-dilation). Range [0,65535] in N \itemmorphRred background estimate using morphological opening (erosion-dilation). Range [0,65535] in N \itemmorphR.erodered background estimate using morphological erosion. Range [0,65535] in N \itemmorphR.close.openred background estimate using morphological closing-opening (dilation-erosion-dilation). Range [0,65535] in N

\item

logratio== log((Rmedian-morphR)/(Gmedian-morphG), base=2), i.e. Redundant. \itemperimeter== 2*sqrt(pi*area/circularity), i.e. Redundant. \itemcircularityShape of spot defined as 4*pi*area/perimeter**2. \itembadspotIf the spot area is greater than product of the horizontal and the vertical average spot separations, equal to 1, otherwise 0.

About IQR

The interquartile range (IQR) is the distance between the 75% quantile (percentile) and the 25% quantile. In words, IQR is the range of the mid 50%. Thus, no outliers are included in the measure, which is why we say it is a robust measure. For norammly distributed data IQR = 1.35*σ, where σ is the standard deviation.

More on background estimates

The Spot software provides several different kinds of background estimates where three of them are based on morphological methods. For all of these methods, the signal selected to be the background signal is the pixel value at the center of the spot after applying the morphological transform using a square mask with side 2.5 times the average distance between two spots. The first and also the simpliest transform (morph.erode) performs a single erosion step. The second transform (morph) performs an opening, which is an erosion followed by a dilution. The third transform (morph.close.open) performs a closing followed by an opening, which is the same as doing a dilution, then an erosion and a dilution again. As the names of the steps indicate, an erosion makes the signal smaller and the dilution the signal larger. Hence, background estimated based on these three methods can always be ordered as morph.erode <= morph <= morph.close.open.

Author(s)

Henrik Bengtsson (http://www.braju.com/R/)

References

Spot Software package by CSIRO, Australia, http://www.cmis.csiro.au/iap/spot.htm

Spot: Description of Output, 2003 http://www.cmis.csiro.au/iap/Spot/spotoutput.htm

Y.H. Yang, M. Buckley, S. Dudoit, T. Speed, Comparison of methods for image analysis on cDNA microarray data, Tech. Report 584, Nov 2000. http://www.stat.berkeley.edu/users/terry/zarray/Html/image.html

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
   spot <- SpotData$read("spot123.spot", path=system.file("data-ex", package="aroma"))

   # Get the foreground and the background (and the layout)
   raw <- getRawData(spot)

   # The the background corrected data
   ma <- getSignal(raw, bgSubtract=FALSE)

   subplots(4, ncol=2)

   # Plot R vs G with a lowess line through the data points
   rg <- as.RGData(ma)
   plot(rg)
   lowessCurve(rg, lwd=2, gridwise=TRUE)

   # Plot M vs A with a lowess line through the data points
   plot(ma)
   lowessCurve(ma, lwd=2, gridwise=TRUE)

   # Plot spatial
   plotSpatial(ma)
 

HenrikBengtsson/aroma documentation built on May 8, 2017, 3:58 p.m.