Preprocessing of cDNAmicroarray DataStatistics 246, Spring 2002,Week 7, Lecture 2Was the experiment a success?What analysis tools should be used?Are there any specific problems?Begin by looking at the dataRed/Green overlay imagesGood: low bg, lots of d.e. Bad: high bg, ghost spots, little d.e.Co-registration and overlay offers a quick visualization,revealing information on color balance, uniformity ofhybridization, spot uniformity, background, and artifiactssuch as dust or scratchesAlways log, always rotatelog2R vs log2G M=log2R/G vs A=log2√√√√RGSignal/Noise = log2(spot intensity/background intensity)HistogramsBoxplots of log2R/G Liver samples from 16 mice: 8 WT, 8 ApoAI KO.Spatial plots: background from the two slidesHighlighting extreme log ratiosTop (black) and bottom (green) 5% of log ratiosBoxplots and highlighting Clear example of spatial biasPrint-tip groupsLog-ratiospin group #Pin group (sub-array) effectsBoxplots of log ratios by pin groupLowess lines through points from pin groupsPlate effectsKO #8Probes: ~6,000 cDNAs, including 200 related to lipid metabolism. Arranged in a 4x4 array of 19x21 sub-arrays.Time of printing effectsGreen channel intensities (log2G). Printing over 4.5 days.The previous slide depicts a slide from this print run.spot numberNormalization Why? To correct for systematic differences betweensamples on the same slide, or between slides,which do not represent true biologicalvariation between samples.How do we know it is necessary? By examining self-self hybridizations, where notrue differential expression is occurring.We find dye biases which vary with overall spotintensity, location on the array, plate origin,pins, scanning parameters,….Self-self hybridizationsFalse color overlay Boxplots within pin-groups Scatter (MA-)plotsFrom the NCI60 data set (Stanford web site)A series of non self-self hybridizationsEarly Ngai lab, UC BerkeleyEarly Goodman lab, UC BerkeleyFrom the Ernest Gallo Clinic & Research CenterEarly PMCRI, Melbourne AustraliaNormalization: methods a) Normalization based on a global adjustment log2 R/G -> log2 R/G - c = log2 R/(kG) Choices for k or c = log2k are c = median or mean of log ratios for aparticular gene set (e.g. housekeeping genes). Or, total intensitynormalization, where k = ∑Ri/ ∑Gi. b) Intensity-dependent normalization. Here we run a line through the middle of the MA plot, shifting the Mvalue of the pair (A,M) by c=c(A), i.e. log2 R/G -> log2 R/G - c (A) = log2 R/(k(A)G). One estimate of c(A) is made using the LOWESS function of Cleveland (1979): LOcally WEighted Scatterplot Smoothing.Normalization: methodsc) Within print-tip group normalization. In addition to intensity-dependent variation in log ratios, spatial biascan also be a significant source of systematic error. Most normalization methods do not correct for spatial effectsproduced by hybridization artifacts or print-tip or plate effectsduring the construction of the microarrays. It is possible to correct for both print-tip and intensity-dependent biasby performing LOWESS fits to the data within print-tip groups, i.e. log2 R/G -> log2 R/G - ci(A) = log2 R/(ki(A)G), where ci(A) is the LOWESS fit to the MA-plot for the ith grid only.Which spots to use for normalization? The LOWESS lines can be run through many different sets ofpoints, and each strategy has its own implicit set of assumptionsjustifying its applicability. For example, we can justify the use of a global LOWESS approachby supposing that, when stratified by mRNA abundance, a) only aminority of genes are expected to be differentially expressed, orb) any differential expression is as likely to be up-regulation asdown-regulation. Pin-group LOWESS requires stronger assumptions: that one ofthe above applies within each pin-group. The use of other sets of genes, e.g. control or housekeepinggenes, involve similar assumptions.Use of control spotsM = log R/G = logR - logG A = ( logR + logG) /2Positive controls(spotted in varying concentrations)NegativecontrolsblanksLowesscurveGlobal scale, global lowess, pin-group lowess; spatial plot after, smooth histograms of M afterMSP titration series(Microarray Sample Pool)Control set to aid intensity- dependent normalizationDifferent concentrationsSpotted evenly spread across the slidePool thewhole libraryYellow: GAPDH, tubulin Light blue: MSP pool / titrationOrange: Schadt-Wong rank invariant set Red line: lowess smooth MSP normalization compared to other methodsComposite normalizationBefore and after compositenormalization-MSP lowess curve-Global lowess curve-Composite lowess curve(Other colours control spots)ci(A)=αAg(A)+(1-αA)fi(A)Comparison of Normalization Schemes(courtesy of Jason Goncalves) No consensus on best segmentation or normalizationmethod Scheme was applied to assess the commonnormalization methods Based on reciprocal labeling experiment data for aseries of 140 replicate experiments on two differentarrays each with 19,200 spotsDESIGN OF RECIPROCALLABELING EXPERIMENT Replicate experiment inwhich we assess thesame mRNA pools butinvert the fluors used. The replicates areindependent experimentsand are scanned,quantified andnormalized as usual2.2/121.2/12)(log)(logExpChChGeneAExpChChGeneARatioRatio −=The following relationship would be observedfor reciprocal microarray experiments inwhich the slides are free of defects and thenormalization scheme performed ideallyWe can measure using real data sets how welleach microarray normalization schemeapproaches this ideal2.2/121.2/12)(log)(logExpChChGeneAExpChChGeneASpotRatioRatioDeviation +=nRatioRatioDeviationExpChChGeneNExpChChGeneNngeArrayAvera2.2/121.2/121)(log)(log +=∑Deviation metric to assessnormalization schemesWe now use the mean array average deviation to compare thenormalization methods. Note that this comparison addresses onlyvariance (precision) and not bias (accuracy) aspects of normalization.Comparison of Normalization Methods - Using 140 19K Microarrays0.30.320.340.360.380.40.420.440.46Pre Normalized Global Intensity Subarray Intensity Global Ratio Sub-Array Ratio Global LOWESS Subarray LOWESSNormalization MethodAverage Mean Deviation Value***Scale normalization: between slidesBoxplots of log ratios from 3 replicate self-self hybridizations.Left panel: before normalizationMiddle panel: after within print-tip group
View Full Document
Unlocking...