| Title: | Optimized local intensity-dependent normalisation of two-color microarrays |
|---|---|
| Description: | Functions for normalisation of two-color microarrays by optimised local regression and for detection of artefacts in microarray data |
| Authors: | Matthias Futschik <[email protected]> |
| Maintainer: | Matthias Futschik <[email protected]> |
| License: | GPL-2 |
| Version: | 1.85.0 |
| Built: | 2024-10-30 08:27:29 UTC |
| Source: | https://github.com/bioc/OLIN |
This function performs an one-factorial analysis of variance assessing intensity-dependent bias for a single array. The predictor variable is the average logged intensity of both channels and the response variable is the logged fold-change.
anovaint(obj,index,N=10)anovaint(obj,index,N=10)
obj |
object of class “marrayRaw” or “marrayNorm” |
index |
index of array to be tested |
N |
number of (intensity) levels for ANOVA |
The function anovaint performs a one-factorial ANOVA for objects of class “marrayRaw” or
“marrayNorm”. The predictor variable is the average logged intensity of both channels
A=0.5*(log2(Ch1)+log2(Ch2)). Ch1,Ch2 are the fluorescence intensities of channel 1
and channel 2, respectively. The response variable is the logged fold-change
M=(log2(Ch2)-log2(Ch1)). The A-scale is divided in N intervals generating
N levels of factor A. Note that
N should divide the total number of spots approx. equally.
The null hypothesis is the equality of mean(M) of the different levels (intervals).
The model formula used is (without an intercept term).
The return value is a list of summary statistics of the fitted model as produced by summary.lm.
For example, the squared multiple correlation coefficient equals the proportion
of the variation of M that can be explained by the variation of A (based on the chosen
ANOVA model.)
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
anova, summary.lm,
anovaspatial, marrayRaw, marrayNorm
# CHECK RAW DATA FOR INTENSITY-DEPENDENT BIAS data(sw) print(anovaint(sw,index=1,N=10)) # CHECK DATA NORMALISED BY OLIN FOR INTENSITY-DEPENDENT BIAS data(sw.olin) print(anovaint(sw.olin,index=1,N=10))# CHECK RAW DATA FOR INTENSITY-DEPENDENT BIAS data(sw) print(anovaint(sw,index=1,N=10)) # CHECK DATA NORMALISED BY OLIN FOR INTENSITY-DEPENDENT BIAS data(sw.olin) print(anovaint(sw.olin,index=1,N=10))
This function performs an one-factorial analysis of variance assessing pin-dependent bias for a single array
anovapin(obj,index)anovapin(obj,index)
obj |
object of class “marrayRaw” or “marrayNorm” |
index |
index of array to be tested |
The function anovapin performs a one-factorial ANOVA for objects of class “marrayRaw” or
“marrayNorm”. The predictor variable is the pin index; the response variable is the logged fold-change
M=(log2(Ch2)-log2(Ch1)).
The null hypothesis is equal mean(M) of groups of spots printed by the same pin i.e.
a spot's M does not dependent on the pin used from printing the spot.
The model formula used is (without an intercept term).
The return value is a list of summary statistics of the fitted model as produced by summary.lm.
For example, the squared multiple correlation coefficient equals the proportion
of the variation of M that can be explained by the variation of pin index (based on the chosen
ANOVA model.)
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
# CHECK RAW DATA FOR INTENSITY-DEPENDENT BIAS data(sw) print(anovapin(sw,index=1)) # CHECK DATA NORMALISED BY OLIN FOR INTENSITY-DEPENDENT BIAS data(sw.olin) print(anovapin(sw.olin,index=1))# CHECK RAW DATA FOR INTENSITY-DEPENDENT BIAS data(sw) print(anovapin(sw,index=1)) # CHECK DATA NORMALISED BY OLIN FOR INTENSITY-DEPENDENT BIAS data(sw.olin) print(anovapin(sw.olin,index=1))
This function performs an one-factorial analysis of variance assessing microtiter plate-dependent bias for a single array
anovaplate(obj,index)anovaplate(obj,index)
obj |
object of class “marrayRaw” or “marrayNorm” |
index |
index of array to be tested |
The function anovapin performs a one-factorial ANOVA for objects of class “marrayRaw” or
“marrayNorm”. The predictor variable is the corresponding plate index as stored in the maPlate slot of obj;
the response variable is the logged fold-change M=(log2(Ch2)-log2(Ch1)).
The null hypothesis is equal mean(M) of groups of spots derived from the same microtiter plate i.e.
a spot's M does not dependent on the plate of origin.
The model formula used is (without an intercept term).
The return value is a list of summary statistics of the fitted model as produced by summary.lm.
For example, the squared multiple correlation coefficient equals the proportion
of the variation of M that can be explained by the variation of plate index (based on the chosen
ANOVA model.)
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
# CHECK RAW DATA FOR INTENSITY-DEPENDENT BIAS data(sw) print(anovapin(sw,index=1)) # CHECK DATA NORMALISED BY OLIN FOR INTENSITY-DEPENDENT BIAS data(sw.olin) print(anovapin(sw.olin,index=1))# CHECK RAW DATA FOR INTENSITY-DEPENDENT BIAS data(sw) print(anovapin(sw,index=1)) # CHECK DATA NORMALISED BY OLIN FOR INTENSITY-DEPENDENT BIAS data(sw.olin) print(anovapin(sw.olin,index=1))
This function performs an one-factorial analysis of variance to test for spatial bias for a single array. The predictor variable is the average logged intensity of both channels and the response variable is the logged fold-change.
anovaspatial(obj,index,xN=5,yN=5,visu=FALSE)anovaspatial(obj,index,xN=5,yN=5,visu=FALSE)
obj |
object of class “marrayRaw” or “marrayNorm” |
index |
index of array (within |
xN |
number of intervals in x-direction |
yN |
number of intervals in y-direction |
visu |
If visu=TRUE, results are visualised (see below) |
The function anovaspatial performs a one-factorial ANOVA for objects of class “marrayRaw” or
“marrayNorm”. The predictor variable is the average logged intensity of both channels
(A=0.5*(log2(Ch1)+log2(Ch2))). Ch1,Ch2 are the fluorescence intensities of channel 1
and channel 2, respectively. The response variable is the logged fold-change
(M=(log2(Ch2)-log2(Ch1))). The spot locations on the array is divided into xN intervals
in x-direction and yN intervals in y-direction. This division defines (xN x yN)
rectangular spatial blocks on
the array, and thus, (xN x yN) levels (or treatments) for A. Note that
values chosen for xN and yN should divide the array columns and rows approx. equally.
The null hypothesis is the equality of mean(M) of the different levels.
The model formula used by anovaspatial is
(without an intercept term).
The return value is a list of summary statistics of the fitted model as produced by summary.lm.
For example, the squared multiple correlation coefficient equals the proportion
of the variation of M that can be related to the spot location (based on the chosen
ANOVA.) Optionally, the distribution of p-values (as derived by t-test and stated in the summary statistics)
can be visualised.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
anova, summary.lm,
anovaint, marrayRaw,
marrayNorm
# CHECK RAW DATA FOR SPATIAL BIAS data(sw) print(anovaspatial(sw,index=1,xN=8,yN=8,visu=TRUE)) # CHECK DATA NORMALISED BY OLIN FOR SPATIAL BIAS data(sw.olin) print(anovaspatial(sw.olin,index=1,xN=8,yN=8,visu=TRUE)) # note the different scale of the colour bar# CHECK RAW DATA FOR SPATIAL BIAS data(sw) print(anovaspatial(sw,index=1,xN=8,yN=8,visu=TRUE)) # CHECK DATA NORMALISED BY OLIN FOR SPATIAL BIAS data(sw.olin) print(anovaspatial(sw.olin,index=1,xN=8,yN=8,visu=TRUE)) # note the different scale of the colour bar
Background correction based on
backgroundCorrect of the limma package.
backgroundCorrect2(object,method="subtract", offset=0)backgroundCorrect2(object,method="subtract", offset=0)
object |
object of class |
method |
method for background correction: “none”, “subtract”, “half”, “minimum”, “movingmin”, “edwards” or “normexp”. |
offset |
numeric value to add to intensities |
This function is a wrapper function for
backgroundCorrect
with following methods implemented:
“none”: no background correction
“subtract”: simple subtraction of background intensities
“movingmin”: background intensities are first averaged over 3x3 grids of neighbouring spots and subsequently substracted
“minimum”: zero or negative intensities after background correction are set equal to half the minimum of positive corrected intensities
“edwards”: background correction based on log-linear interpolation
“normexp”: background correction based on fitting procedure
For further details and references, please refer to its help page. An
alternative Bayesian model for background correction
(kooperberg) is also implemented
in the limma package.
Background correct object of class marrayRaw.
Matthias Futschik
# Loading data data(sw) #No background correction sw.none <- backgroundCorrect2(sw,method="none") plot(maA(sw.none)[,1],maM(sw.none)[,1]) # Simple subtraction sw.sub <- backgroundCorrect2(sw,method="sub") points(maA(sw.sub)[,1],maM(sw.sub)[,1],col="red")# Loading data data(sw) #No background correction sw.none <- backgroundCorrect2(sw,method="none") plot(maA(sw.none)[,1],maM(sw.none)[,1]) # Simple subtraction sw.sub <- backgroundCorrect2(sw,method="sub") points(maA(sw.sub)[,1],maM(sw.sub)[,1],col="red")
This function performs an between-array scaling
bas(obj,mode="var")bas(obj,mode="var")
obj |
object of “marrayNorm” |
mode |
mode of scaling. Default option is scaling of arrays
to have the same within-array variance of
logged ratios ( |
The function bsv adjust the scale of logged ratios (M=(log2(Ch2)-log2(Ch1)))
between the different arrays stored in obj.
Following schemes (mode) are implemented:
mode="var": Logged ratios M are scaled to show the same (within-array)
variance for all arrays in the batch stored in obj.
The variance is calculated using var.
mode="mad": The same procedure as for mode="var" is applied using, however,
median absolute deviation (mad) as robust estimate for withing-array variance.
mode="qq": The quantile scaling is using the same procedure as the quantile normalisation described
by Bolstad et al. (2003). In brief: Given X is the matrix with logged ratios (column corresponding to arrays, rows to genes)
Sort each column of X (independently) producing Xs,
Replace values in each row of Xs by the mean value of the row producing Xsm,
Rearrange the ordering for each column of matrix Xsm, so that it has the columns have same ordering as for the original matrix X.
The last step yields the scaled logged ratios M.
Between-array scaling should only be performed if it can be assumed that the different arrays have a similar distribution of logged ratios. This has to be check on a case-by-case basis. Caution should be taken in the interpretation of results for arrays hybridised with biologically divergent samples, if between-array scaling is applied.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
Bolstad et al., A comparison of normalization methods for high density oligonucleotide array data based on variance and bias, Bioinformatics, 19: 185-193, 2003
# DISTRIBUTION OF M BEFORE SCALING data(sw.olin) col <- c("red","blue","green","orange") M <- maM(sw.olin) plot(density(M[,4]),col=col[4],xlim=c(-2,2)) for (i in 1:3){ lines(density(M[,i]),col=col[i]) } # SCALING AND VISUALISATION sw.olin.s <- bas(sw.olin,mode="var") M <- maM(sw.olin.s) plot(density(M[,4]),col=col[4],xlim=c(-2,2)) for (i in 1:3){ lines(density(M[,i]),col=col[i]) }# DISTRIBUTION OF M BEFORE SCALING data(sw.olin) col <- c("red","blue","green","orange") M <- maM(sw.olin) plot(density(M[,4]),col=col[4],xlim=c(-2,2)) for (i in 1:3){ lines(density(M[,i]),col=col[i]) } # SCALING AND VISUALISATION sw.olin.s <- bas(sw.olin,mode="var") M <- maM(sw.olin.s) plot(density(M[,4]),col=col[4],xlim=c(-2,2)) for (i in 1:3){ lines(density(M[,i]),col=col[i]) }
Generates colour bar for MXY plots
colorbar.mxy(color.lim, col=c(rgb(0,(100:0)/100,0),rgb(0,0,0),rgb((1:100)/100,green=0,blue=0)), ylab="",ylablim=FALSE)colorbar.mxy(color.lim, col=c(rgb(0,(100:0)/100,0),rgb(0,0,0),rgb((1:100)/100,green=0,blue=0)), ylab="",ylablim=FALSE)
color.lim |
limits for colour range |
col |
colour palette to be used |
ylab |
label of ordinate of color bar |
ylablim |
If TRUE, the axis annotation consists only of the limits of the colour range. |
The function colorbar.mxy produces a colour bar for MXY plots.
The default colours used range from green (for the lower limit of the colour range)
to red (for its upper limit).
For visualisation, values below or above the limits for the colour range
(as given by color.lim) are set
to the lower or upper limit, respectively.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
data(sw) # GENERATING LAYOUT mat <- matrix(1:2,ncol=2,nrow=1,byrow=TRUE) l <- layout(mat,widths=c(5,1)) # CHOOSING LIMITS OF COLOUR RANGE color.lim <- c(-2,2) # PLOTTING Mtmp <- v2m(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw),Nsc=maNsc(sw),Nsr=maNsr(sw), visu=TRUE,color.lim=color.lim) colorbar.mxy(color.lim=color.lim,ylablim=FALSE,ylab="M")data(sw) # GENERATING LAYOUT mat <- matrix(1:2,ncol=2,nrow=1,byrow=TRUE) l <- layout(mat,widths=c(5,1)) # CHOOSING LIMITS OF COLOUR RANGE color.lim <- c(-2,2) # PLOTTING Mtmp <- v2m(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw),Nsc=maNsc(sw),Nsr=maNsr(sw), visu=TRUE,color.lim=color.lim) colorbar.mxy(color.lim=color.lim,ylablim=FALSE,ylab="M")
Generates colour bar for 2D plots of absolute values
colorbar.mxy.abs(color.lim,color="red", ylab="",ylablim=FALSE)colorbar.mxy.abs(color.lim,color="red", ylab="",ylablim=FALSE)
color.lim |
limits for colour range |
color |
colour to be used: “red” or “green” |
ylab |
label of y-axis |
ylablim |
If TRUE, the axis annotation consists only of the limits of the colour range. |
The function colorbar.mxy.abs is a modification of colorbar.mxy
to provide colour-bars for MXY plots of absolutes values. It is used in function
mxy.abs.plot.
Further details can be found at colorbar.mxy.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
mxy.abs.plot, colorbar.mxy, colorbar.sig
This function generates a colour bar for the visualisation of significance of spatial bias.
colorbar.sig(color.lim=c(-3,3))colorbar.sig(color.lim=c(-3,3))
color.lim |
limits for color bar |
The function colorbar.sig produces a colour bar for 2D-plots generated by sigxy.plot.
The colours used range from green (for the lower limit of the colour range)
to red (for its upper limit).
For visualisation, values below or above the limits for the colour range
(as given by color.lim) are set
to the lower or upper limit, respectively. The function colorbar.sig
is similar to more general function colorbar.mxy.
It differs, however, in its axis annotation. Since it is used to present the significance
in a log10-scale, annotation of axis tacks consists of negative values in both direction.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
This function assesses the significance of intensity-dependent bias by an one-sided random permutation test. The observed average values of logged fold-changes within an intensity neighbourhood are compared to an empirical distribution generated by random permutation. The significance is given by the false discovery rate.
fdr.int(A,M,delta=50,N=100,av="median")fdr.int(A,M,delta=50,N=100,av="median")
A |
vector of average logged spot intensity |
M |
vector of logged fold changes |
delta |
integer determining the size of the neighbourhood. The actual window size is
( |
N |
number of random permutations performed for generation of empirical distribution |
av |
averaging of |
The function fdr.int assesses significance of intensity-dependent bias using a one-sided random permutation test.
The null hypothesis states the independence of A and M. To test if M depends on A,
spots are ordered with respect to A. This defines a neighbourhood of spots with similar A for each spot.
Next, a test statistic is defined by calculating the median or mean of M
within
a symmetrical spot's intensity neighbourhood of chosen size (2 *delta+1). An empirical distribution of the
test statistic is produced by calculating for N random intensity orders of spots.
Comparing this empirical distribution of
with the observed distribution of ,
the independence of M and A
is assessed. If M is independent of A, the empirical distribution
of can be expected to be
distributed around its mean value. The false discovery rate (FDR) is used to
assess the significance of observing positive deviations of .
It indicates the expected proportion of false positives
among rejected null hypotheses. It is defined as ,
where q is the fraction of larger than chosen threshold c
for the empirical distribution, s is the number of neighbourhoods with
for the distribution derived from the original data and T
is the total number of neighbourhoods in the original data.
Varying threshold c determines the FDR for each spot neighbourhood. FDRs equal zero are set to
for computational reasons, as log10(FDR) is plotted by sigint.plot.
Correspondingly, the significance
of observing negative deviations of can be determined. If the neighbourhood
window extends over the limits of the intensity scale, the significance is set to NA.
A list of vector containing the false discovery rates for positive (FDRp) and negative (FDRn) deviations of
(of the spot's neighbourhood) is produced.
The same functionality but with our input and output formats is offered by fdr.int
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
fdr.int2,p.int, fdr.spatial, sigint.plot
# To run these examples, delete the comment signs (#) in front of the commands. # # LOADING DATA NOT-NORMALISED # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this example, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.int(maA(sw)[,1],maM(sw)[,1],delta=50,N=10,av="median") # VISUALISATION OF RESULTS # sigint.plot(maA(sw)[,1],maM(sw)[,1],FDR$FDRp,FDR$FDRn,c(-5,-5)) # LOADING NORMALISED DATA # data(sw.olin) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # FDR <- fdr.int(maA(sw.olin)[,1],maM(sw.olin)[,1],delta=50,N=10,av="median") # VISUALISATION OF RESULTS # sigint.plot(maA(sw.olin)[,1],maM(sw.olin)[,1],FDR$FDRp,FDR$FDRn,c(-5,-5))# To run these examples, delete the comment signs (#) in front of the commands. # # LOADING DATA NOT-NORMALISED # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this example, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.int(maA(sw)[,1],maM(sw)[,1],delta=50,N=10,av="median") # VISUALISATION OF RESULTS # sigint.plot(maA(sw)[,1],maM(sw)[,1],FDR$FDRp,FDR$FDRn,c(-5,-5)) # LOADING NORMALISED DATA # data(sw.olin) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # FDR <- fdr.int(maA(sw.olin)[,1],maM(sw.olin)[,1],delta=50,N=10,av="median") # VISUALISATION OF RESULTS # sigint.plot(maA(sw.olin)[,1],maM(sw.olin)[,1],FDR$FDRp,FDR$FDRn,c(-5,-5))
This function assesses the significance of intensity-dependent bias by an one-sided random permutation test. The observed average values of logged fold-changes within an intensity neighbourhood are compared to an empirical distribution generated by random permutation. The significance is given by the false discovery rate.
fdr.int2(object,delta=50,N=100,av="median")fdr.int2(object,delta=50,N=100,av="median")
object |
object of class marrayRaw or marrayNorm |
delta |
integer determining the size of the neighbourhood. The actual window size is
( |
N |
number of random permutations performed for generation of empirical distribution |
av |
averaging of |
This function fdr.int2 is basically the same as fdr.int except for
differences in their in- and output format. For the details about the functionality, see fdr.int.
This function will be merged with fdr.int in future versions.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
# To run these examples, delete the comment signs (#) in front of the commands. # # LOADING DATA NOT-NORMALISED # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this example, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.int2(sw,delta=50,N=10,av="median") # VISUALISATION OF RESULTS # sigint.plot2(sw[,1],FDR$FDRp[[1]],FDR$FDRn[[1]],c(-5,-5)) # array 1 # sigint.plot2(sw[,4],FDR$FDRp[[4]],FDR$FDRn[[4]],c(-5,-5)) # array 4# To run these examples, delete the comment signs (#) in front of the commands. # # LOADING DATA NOT-NORMALISED # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this example, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.int2(sw,delta=50,N=10,av="median") # VISUALISATION OF RESULTS # sigint.plot2(sw[,1],FDR$FDRp[[1]],FDR$FDRn[[1]],c(-5,-5)) # array 1 # sigint.plot2(sw[,4],FDR$FDRp[[4]],FDR$FDRn[[4]],c(-5,-5)) # array 4
This function assesses the significance of spatial bias by a one-sided random permutation test. This is achieved by comparing the observed average values of logged fold-changes within a spot's spatial neighbourhood with an empirical distribution generated by random permutation. The significance of spatial bias is given by the false discovery rate.
fdr.spatial(X,delta=2,N=100,av="median",edgeNA=FALSE)fdr.spatial(X,delta=2,N=100,av="median",edgeNA=FALSE)
X |
matrix of logged fold changes.
For alternative input format, see |
delta |
integer determining the size of spot neighbourhoods
( |
N |
number of random permutations performed for generation of empirical background distribution |
av |
averaging of |
edgeNA |
treatment of edges of array: For |
The function fdr.spatial assesses the significance of spatial bias using a one-sided random permutation test.
The null hypothesis states random spotting i.e. the independence of log ratio M
and spot location. First, a neighbourhood of a spot is defined by a two dimensional square window
of chosen size ((2*delta+1)x(2*delta+1)). Next, a test statistic is defined by calculating
the median or mean of M within
a symmetrical spot's neighbourhood. An empirical distribution of is generated
based
N random permutations of the spot locations on the array. The randomisation and calculation of
is repeated N times.
Comparing this empirical distribution of
with the observed distribution of ,
the independence of M and spot location
can be assessed. If M is independent of spot's location,
the empirical distribution can be expected to be
distributed around its mean value. To assess the significance of observing positive deviations of
,
the false discovery rate (FDR) is used. It indicates the expected proportion of false discoveries
among rejected null hypotheses. It is defined as ,
where q is the fraction of larger than chosen threshold c
for the empirical distribution, s is the number of neighbourhoods with
for the distribution derived from the original data and T
is the total number of neighbourhoods on the array. FDRs equal zero are set to
.
Varying threshold c determines the FDR for each spot neighbourhood. Correspondingly, the significance
of observing negative deviations of can be determined.
A list of matrices containing the false discovery rates for positive (FDRp)
and negative (FDRn) deviations of
of the spot's neighbourhood is produced.
The same functionality but with our input and output formats is offered by fdr.spatial
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
p.spatial, fdr.int, sigxy.plot, fdr.spatial2
# To run these examples, delete the comment signs before the commands. # # LOADING DATA # data(sw) # M <- v2m(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw), # Nsc=maNsc(sw),Nsr=maNsr(sw),main="MXY plot of SW-array 1") # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.spatial(M,delta=2,N=10,av="median",edgeNA=TRUE) # sigxy.plot(FDR$FDRp,FDR$FDRn,color.lim=c(-5,5),main="FDR") # # LOADING NORMALISED DATA # data(sw.olin) # M<- v2m(maM(sw.olin)[,1],Ngc=maNgc(sw.olin),Ngr=maNgr(sw.olin), # Nsc=maNsc(sw.olin),Nsr=maNsr(sw.olin),main="MXY plot of SW-array 1") # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # FDR <- fdr.spatial(M,delta=2,N=10,av="median",edgeNA=TRUE) # VISUALISATION OF RESULTS # sigxy.plot(FDR$FDRp,FDR$FDRn,color.lim=c(-5,5),main="FDR")# To run these examples, delete the comment signs before the commands. # # LOADING DATA # data(sw) # M <- v2m(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw), # Nsc=maNsc(sw),Nsr=maNsr(sw),main="MXY plot of SW-array 1") # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.spatial(M,delta=2,N=10,av="median",edgeNA=TRUE) # sigxy.plot(FDR$FDRp,FDR$FDRn,color.lim=c(-5,5),main="FDR") # # LOADING NORMALISED DATA # data(sw.olin) # M<- v2m(maM(sw.olin)[,1],Ngc=maNgc(sw.olin),Ngr=maNgr(sw.olin), # Nsc=maNsc(sw.olin),Nsr=maNsr(sw.olin),main="MXY plot of SW-array 1") # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # FDR <- fdr.spatial(M,delta=2,N=10,av="median",edgeNA=TRUE) # VISUALISATION OF RESULTS # sigxy.plot(FDR$FDRp,FDR$FDRn,color.lim=c(-5,5),main="FDR")
This function assesses the significance of spatial bias by a one-sided random permutation test. This is achieved by comparing the observed average values of logged fold-changes within a spot's spatial neighbourhood with an empirical distribution generated by random permutation. The significance of spatial bias is given by the false discovery rate.
fdr.spatial2(object,delta=2,N=100,av="median",edgeNA=FALSE)fdr.spatial2(object,delta=2,N=100,av="median",edgeNA=FALSE)
object |
object of class marrayRaw or marrayNorm |
delta |
integer determining the size of spot neighbourhoods
( |
N |
number of random permutations performed for generation of empirical background distribution |
av |
averaging of |
edgeNA |
treatment of edges of array: For |
The function fdr.spatial2.Rd is basically the same as fdr.spatial,
but differs in its input and output formats.
Details about the functionality can be found
at fdr.spatial.
Two list of vectors containing the false discovery rates for positive (FDRp)
and negative (FDRn) deviations of
of the spot's neighbourhood is produced. Each
vector contains the false discovery values for one array.
This function will be fused with fdr.spatial in future versions using S4-style methods.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
p.spatial, fdr.int, sigxy.plot,
# To run these examples, delete the comment signs before the commands. # # LOADING DATA # data(sw) # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.spatial2(sw,delta=2,N=10,av="median",edgeNA=TRUE) # # SIGNIFICANCE PLOTS OF ARRAY 1 # sigxy.plot2(sw[,1],FDR$FDRp[[1]],FDR$FDRn[[1]],color.lim=c(-5,5),main="FDR") # SIGNIFICANCE PLOTS OF ARRAY 3 # sigxy.plot2(sw[,3],FDR$FDRp[[3]],FDR$FDRn[[3]],color.lim=c(-5,5),main="FDR") ## To run these examples, delete the comment signs before the commands. # # LOADING DATA # data(sw) # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.spatial2(sw,delta=2,N=10,av="median",edgeNA=TRUE) # # SIGNIFICANCE PLOTS OF ARRAY 1 # sigxy.plot2(sw[,1],FDR$FDRp[[1]],FDR$FDRn[[1]],color.lim=c(-5,5),main="FDR") # SIGNIFICANCE PLOTS OF ARRAY 3 # sigxy.plot2(sw[,3],FDR$FDRp[[3]],FDR$FDRn[[3]],color.lim=c(-5,5),main="FDR") #
This function generates 2D-plots of the foreground, background and background corrected fluorescence intensities of channel 1 and of channel 2, respectively.
fgbg.visu(obj,label)fgbg.visu(obj,label)
obj |
object of class “marrayRaw” |
label |
character string for labelling. It will be added to the title of the first sub-plot. |
The function fgbg.visu produces 2D-representations of the foreground
and background intensities for both fluorescence channels (as stored in
obj). Additionally, a plot of the difference between fore- and
background intensities is generated (background-corrected intensities).
All intensities are log2-transformed. The colour range for plotting is defined by 0 and the maximum of
the logged intensity for each sub-graph separately.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
# LOADING RAW DATA data(sw) # PLOTTING fgbg.visu(sw[,1])# LOADING RAW DATA data(sw) # PLOTTING fgbg.visu(sw[,1])
This functions performs intensity-dependent normalisation based on local regression by locfit.
ino(object,alpha=0.3,weights=NA,bg.corr="subtract",...)ino(object,alpha=0.3,weights=NA,bg.corr="subtract",...)
object |
object of class “marrayRaw” or “marrayNorm” |
alpha |
smoothing parameter |
weights |
matrix of weights for local regression. Rows correspond to the spotted probe sequences, columns to arrays in the batch. These may be derived from the matrix of spot quality weights as defined for “maRaw” objects. |
bg.corr |
backcorrection method (for “marrayRaw” objects) : “none” or “subtract”(default). |
... |
Further arguments for |
The function ino regresses the average logged fold changes (M) with respect to the average
logged spot intensity (A). The residuals of this fit are the normalised logged fold changes.
The parameter alpha
specifies the fraction of points that are included in the neighbourhood and thus has a value between 0 and 1.
Larger alpha values lead to smoother fits.
Object of class “marrayNorm” with normalised logged ratios
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
maNorm, locfit.raw,olin, oin, lin
# LOADING DATA data(sw) # INTENSITY-DEPENDENT NORMALISATION norm.ino <- ino(sw) # MA-PLOT OF NORMALISATION RESULTS OF FIRST ARRAY plot(maA(norm.ino)[,1],maM(norm.ino)[,1],main="INO") # CORRESPONDING MXY-PLOT mxy.plot(maM(norm.ino)[,1],Ngc=maNgc(norm.ino),Ngr=maNgr(norm.ino), Nsc=maNsc(norm.ino),Nsr=maNsr(norm.ino),main="INO")# LOADING DATA data(sw) # INTENSITY-DEPENDENT NORMALISATION norm.ino <- ino(sw) # MA-PLOT OF NORMALISATION RESULTS OF FIRST ARRAY plot(maA(norm.ino)[,1],maM(norm.ino)[,1],main="INO") # CORRESPONDING MXY-PLOT mxy.plot(maM(norm.ino)[,1],Ngc=maNgc(norm.ino),Ngr=maNgr(norm.ino), Nsc=maNsc(norm.ino),Nsr=maNsr(norm.ino),main="INO")
This functions performs local intensity-dependent normalisation (LIN)
lin(object,X=NA,Y=NA,alpha=0.3,iter=2,scale=TRUE,weights=NA,bg.corr="subtract",...)lin(object,X=NA,Y=NA,alpha=0.3,iter=2,scale=TRUE,weights=NA,bg.corr="subtract",...)
object |
object of class “marrayRaw” |
X |
matrix with x-coordinates of spots. If X=NA, columns on array are used as proxies for the location in x-direction |
Y |
matrix with y-coordinates of spots. If Y=NA, rows on array are used as proxies for the location in y-direction |
alpha |
smooting parameter for local regression |
iter |
number of iterations in the LIN procedure |
scale |
scale parameter for smooting in Y-direction of the array in respect to smoothing in
X-direcction. If |
weights |
matrix of weights for local regression. Rows correspond to the spotted probe sequences, columns to arrays in the batch. These may be derived from the matrix of spot quality weights as defined for “maRaw” objects. |
bg.corr |
backcorrection method (for “marrayRaw” objects) : “none” or “subtract”(default). |
... |
Further arguments for |
LIN is based on the same normalisation scheme as OLIN, but does not incorporate
optimisation of model parameters. The function lin can serve for comparison.
Alternatively, it can be used to enforce a conservative model fit.
The smoothing parameter alpha controls the neighbourhood size h of local fitting.
It
specifies the fraction of points that are included in the neighbourhood and, thus, has a value between 0 and 1.
Larger alpha values lead to smoother fits.
If the normalisation should be based on set of genes assumed to be not differentially expressed (house-keeping genes), weights can be used for local regression. In this case, all weights should be set to zero except for the house-keeping genes for which weights are set to one. In order to achieve a reliable regression, it is important, however, that there is a sufficient number of house-keeping genes that cover the whole expression range and are spotted accross the whole array.
Object of class “marrayNorm” with normalised logged ratios
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
M.Futschik and T.Crompton (2004) Model selection and efficiency testing for normalization of cDNA microarray data, Genome Biology, 5:R60
# LOADING DATA data(sw) data(sw.xy) # LOCAL INTENSITY-DEPENDENT NORMALISATION norm.lin <- lin(sw,X=sw.xy$X,Y=sw.xy$Y) # MA-PLOT OF NORMALISATION RESULTS OF FIRST ARRAY plot(maA(norm.lin)[,1],maM(norm.lin)[,1],main="LIN") # CORRESPONDING MXY-PLOT mxy.plot(maM(norm.lin)[,1],Ngc=maNgc(norm.lin),Ngr=maNgr(norm.lin), Nsc=maNsc(norm.lin),Nsr=maNsr(norm.lin),main="LIN")# LOADING DATA data(sw) data(sw.xy) # LOCAL INTENSITY-DEPENDENT NORMALISATION norm.lin <- lin(sw,X=sw.xy$X,Y=sw.xy$Y) # MA-PLOT OF NORMALISATION RESULTS OF FIRST ARRAY plot(maA(norm.lin)[,1],maM(norm.lin)[,1],main="LIN") # CORRESPONDING MXY-PLOT mxy.plot(maM(norm.lin)[,1],Ngc=maNgc(norm.lin),Ngr=maNgr(norm.lin), Nsc=maNsc(norm.lin),Nsr=maNsr(norm.lin),main="LIN")
This function converts a matrix based on the spatial layout of spots to a vector. Optionally, a 2D-plot is produced.
m2v(M,Ngc,Ngr,Nsc,Nsr,visu=FALSE,color.lim=c(-1,1),xlab="Columns",ylab="Rows",...)m2v(M,Ngc,Ngr,Nsc,Nsr,visu=FALSE,color.lim=c(-1,1),xlab="Columns",ylab="Rows",...)
M |
Matrix of real values |
Ngc |
number of columns for the grid matrix |
Ngr |
number of rows for the grid matrix |
Nsc |
number of columns for the spot matrix |
Nsr |
number of rows for the spot matrix |
visu |
If TRUE, MXY plot is generated. |
color.lim |
limits of colour range for 2D-plot |
xlab |
label of x -axis of 2D-plot |
ylab |
label of y-axis of 2D-plot |
... |
Further optional parameters for the |
The function m2v rearranges the values of a matrix M corresponding to the intensity values
on the array to a vector V. The matrix M may have been generated by e.g.
v2m. The order of values in V follows the convention of marrayRaw objects.
In fact, the transformation of m2v is the reverse of v2m (assuming the arguments
are kept the same.)
Note that these functions assume a specific mapping between the data points and the location of spot (i.e.
the same mapping rule that is used for marrayRaw/marrayNorm objects.) The validity of the mappings should be
carefully checked (see also the documentation of the marray package.)
The option for spatial visualisation
is rather restricted to logged fold-changes as the corresponding colour range is centred around zero and
follows the conventional colouring (green for negative, red for positive fold-changes). The MXY plot produced
by m2v does not include a colour-bar. To have a colour included, mxy.plot can be used.
A vector V of length ((Ngc*Nsc)*(Ngr*Nsr)) is produced. The values of V represents the spatial
distribution of the values of vector V given the print-layout. Optionally, a 2D-plot of M is generated.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
# LOADING DATA NOT-NORMALISED data(sw) # CONVERSION FROM VECTOR TO MATRIX M <- v2m(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw),Nsc=maNsc(sw),Nsr=maNsr(sw),visu=TRUE) # BACK-CONVERSION FROM MATRIX TO VECTOR V <- m2v(M,Ngc=maNgc(sw),Ngr=maNgr(sw),Nsc=maNsc(sw),Nsr=maNsr(sw),visu=TRUE)# LOADING DATA NOT-NORMALISED data(sw) # CONVERSION FROM VECTOR TO MATRIX M <- v2m(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw),Nsc=maNsc(sw),Nsr=maNsr(sw),visu=TRUE) # BACK-CONVERSION FROM MATRIX TO VECTOR V <- m2v(M,Ngc=maNgc(sw),Ngr=maNgr(sw),Nsc=maNsc(sw),Nsr=maNsr(sw),visu=TRUE)
Using a sliding square window this function produces the moving average for a matrix.
ma.matrix(X,av="median",delta= 2,edgeNA=FALSE )ma.matrix(X,av="median",delta= 2,edgeNA=FALSE )
X |
matrix |
av |
averaging by mean or median (default) |
delta |
integer determining the size of the sliding square window (2*delta+1)x(2*delta+1). |
edgeNA |
treatment of edges of array: For |
A square window with size (2*delta+1)x(2*delta+1) is moved over the entire matrix and a new matrix is created with each value equals the average value in the corresponding window. This procedure defines a local regression of zeroth order.
Matrix with average values of matrix X.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
ma.vector
### LOADING DATA data(sw) ### GENERATION OF MATRIX Morig <- v2m(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw),Nsc=maNsc(sw),Nsr=maNsr(sw),visu=TRUE) ### AVERAGING BY MA.MATRIX Mav <- ma.matrix(Morig,av="median",delta= 2,edgeNA=FALSE ) ### VISUALISATION m2v(Mav,Ngc=maNgc(sw),Ngr=maNgr(sw),Nsc=maNsc(sw),Nsr=maNsr(sw),visu=TRUE)### LOADING DATA data(sw) ### GENERATION OF MATRIX Morig <- v2m(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw),Nsc=maNsc(sw),Nsr=maNsr(sw),visu=TRUE) ### AVERAGING BY MA.MATRIX Mav <- ma.matrix(Morig,av="median",delta= 2,edgeNA=FALSE ) ### VISUALISATION m2v(Mav,Ngc=maNgc(sw),Ngr=maNgr(sw),Nsc=maNsc(sw),Nsr=maNsr(sw),visu=TRUE)
This functions calculates the moving average for a vector.
ma.vector(A,M,av="median",delta=50)ma.vector(A,M,av="median",delta=50)
A |
vector of predictor to be used for sorting |
M |
vector of variable to be averaged |
av |
averaging by mean or median (default) |
delta |
even integer determining the size of the sliding window ( |
The function ma.vector first sorts M according to the corresponding values of A.
Subsequently, a moving average is calculated with window size (2*delta+1). The
values for the moving average are set to zero if the corresponding window extends
over the boarder of the vector M.
Vector with moving average values of M
Matthias E. Futschik,http://itb.biologie.hu-berlin.de/~futschik
### LOADING DATA data(sw) A <- maA(sw[,1]) M <- maM(sw[,1]) # MA-PLOT plot(A,M) # MOVING AVERAGE Mav <- ma.vector(A,M,av="median",delta=100) points(A,Mav,col="red")### LOADING DATA data(sw) A <- maA(sw[,1]) M <- maM(sw[,1]) # MA-PLOT plot(A,M) # MOVING AVERAGE Mav <- ma.vector(A,M,av="median",delta=100) points(A,Mav,col="red")
This function produce a MXY plot of absolute values of M including a colour bar.
mxy.abs.plot(V,Ngc,Ngr,Nsc,Nsr,color.lim,color="red",xlab="Columns",ylab="Rows",...)mxy.abs.plot(V,Ngc,Ngr,Nsc,Nsr,color.lim,color="red",xlab="Columns",ylab="Rows",...)
V |
vector of positive values |
Ngc |
number of columns for the grid matrix |
Ngr |
number of rows for the grid matrix |
Nsc |
number of columns for the spot matrix |
Nsr |
number of rows for the spot matrix |
color.lim |
limits of color range for MXY plot |
color |
color to be used for plot: “red” (default) or “green” |
xlab |
label of x -axis of MXY plot |
ylab |
label of y-axis of MXY plot |
... |
Further optional graphical parameter for the |
The function mxy.abs.plot is similar to function mxy.plot. Details can therefore
be found at mxy.plot. Two differences, however, exist: First, mxy.abs.plot
plots the absolute value of V and second, “red” (default) or “green” can be chosen
as colour of plotting. Hence, mxy.abs.plot facilitates the inspection of
spatial artifacts in single fluorescence channels.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
v2m, m2v, colorbar.mxy.abs, fgbg.visu, image
# LOADING DATA data(sw) # PLOTTING OF ABSOLUTE LOGGED FOLD-CHANGES mxy.plot(abs(maM(sw)[,1]),Ngc=maNgc(sw),Ngr=maNgr(sw),Nsc=maNsc(sw),Nsr=maNsr(sw)) # PLOTTING SPATIAL DISTRIBUTION OF SINGLE-CHANNEL INTENSITIES mxy.abs.plot(maRf(sw)[,1],color.lim=c(0,10000),Ngc=maNgc(sw),Ngr=maNgr(sw), Nsc=maNsc(sw),Nsr=maNsr(sw)) mxy.abs.plot(maGf(sw)[,1],color.lim=c(0,10000),color="green",Ngc=maNgc(sw),Ngr=maNgr(sw), Nsc=maNsc(sw),Nsr=maNsr(sw))# LOADING DATA data(sw) # PLOTTING OF ABSOLUTE LOGGED FOLD-CHANGES mxy.plot(abs(maM(sw)[,1]),Ngc=maNgc(sw),Ngr=maNgr(sw),Nsc=maNsc(sw),Nsr=maNsr(sw)) # PLOTTING SPATIAL DISTRIBUTION OF SINGLE-CHANNEL INTENSITIES mxy.abs.plot(maRf(sw)[,1],color.lim=c(0,10000),Ngc=maNgc(sw),Ngr=maNgr(sw), Nsc=maNsc(sw),Nsr=maNsr(sw)) mxy.abs.plot(maGf(sw)[,1],color.lim=c(0,10000),color="green",Ngc=maNgc(sw),Ngr=maNgr(sw), Nsc=maNsc(sw),Nsr=maNsr(sw))
This function produce a MXY plot including a colour bar.
mxy.plot(V,Ngc,Ngr,Nsc,Nsr,color.lim=c(-1,1),xlab="Columns",ylab="Rows",...)mxy.plot(V,Ngc,Ngr,Nsc,Nsr,color.lim=c(-1,1),xlab="Columns",ylab="Rows",...)
V |
vector of real values typically logged ratios M. Alternatively, V can be an object of class marrayRaw or marrayNorm. In this case, the layout of the array does not need to be given. |
Ngc |
number of columns for the grid matrix |
Ngr |
number of rows for the grid matrix |
Nsc |
number of columns for the spot matrix |
Nsr |
number of rows for the spot matrix |
color.lim |
limits of color range for MXY plot |
xlab |
label of x -axis of MXY plot |
ylab |
label of y-axis of MXY plot |
... |
Further optional graphical parameter for the |
Spotted microarrays have generally a grid layout of form with
Ngc columns and Ngr rows. Each block (or spot matrix)
of the grid corresponds to a specific pin used for spotting.
The blocks have generally Nsc columns and Nsr rows.
The function mxy.plot generates a 2D-plot (MXY-plot) of the values of M across the
array. M is given in form of the vector V.
Note that this function assumes a specific mapping between the data points and the location of spot (i.e.
the same mapping rule that is used for marrayRaw/marrayNorm objects (see the documentation of packet marray)
The colour range of the MXY plot is centred around zero and
follows the conventional colouring (green for negative, red for positive fold-changes). For a separate visualisation\
of the two channels, see function fgbg.visu.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
v2m, m2v, fgbg.visu, image, marrayRaw
# LOADING DATA data(sw) # PLOTTING mxy.plot(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw), Nsc=maNsc(sw),Nsr=maNsr(sw)) # ALTERNATIVE mxy.plot(sw[,1])# LOADING DATA data(sw) # PLOTTING mxy.plot(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw), Nsc=maNsc(sw),Nsr=maNsr(sw)) # ALTERNATIVE mxy.plot(sw[,1])
This function produce a MXY plot with a colour bar.
In contrast to mxy.plot, the plot is based on spot coordinates (instead on column and row
index as proxies for spot location).
mxy2.plot(V,X,Y,Ngc,Ngr,Nsc,Nsr,color.lim=c(-1,1),xlab="X",ylab="Y",...)mxy2.plot(V,X,Y,Ngc,Ngr,Nsc,Nsr,color.lim=c(-1,1),xlab="X",ylab="Y",...)
V |
vector of real values typically logged ratios M. |
X |
vector of x coordinates of spot locations |
Y |
vector of y coordinates of spot locations |
Ngc |
number of columns for the grid matrix |
Ngr |
number of rows for the grid matrix |
Nsc |
number of columns for the spot matrix |
Nsr |
number of rows for the spot matrix |
color.lim |
limits of color range for MXY plot |
xlab |
label of x-axis of MXY plot |
ylab |
label of y-axis of MXY plot |
... |
Further optional graphical parameter for the |
The function mxy2.plot can be used to plot the distribution of V across
the array. As mxy.plot, it mainly aims for the plotting of the distribution of
logged fold changes. It differs from mxy.plot in the representation of spot location.
The function mxy.plot uses the index of columns and rows as proxies for the spot location.
The gaps between the grid matrices (spotted by different pins) are, therefore, not reproduced
in the plot. A more accurate spatial plot is produced by mxy2.plot, which
is based on the coordinates of the first column and first raw of the array. Assuming a regular rectangular print layout,
gaps and the edges of the array are shown.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
mxy.plot, v2m, m2v, fgbg.visu, image
# LOADING DATA data(sw) data(sw.xy) # PLOTTING mxy2.plot(maM(sw)[,1],X=sw.xy$X[,1],Y=sw.xy$Y[,1], Ngc=maNgc(sw),Ngr=maNgr(sw), Nsc=maNsc(sw),Nsr=maNsr(sw))# LOADING DATA data(sw) data(sw.xy) # PLOTTING mxy2.plot(maM(sw)[,1],X=sw.xy$X[,1],Y=sw.xy$Y[,1], Ngc=maNgc(sw),Ngr=maNgr(sw), Nsc=maNsc(sw),Nsr=maNsr(sw))
This functions performs optimised intensity-dependent normalisation (OLIN).
oin(object,alpha=seq(0.1,1,0.1),weights=NA,bg.corr="subtract",...)oin(object,alpha=seq(0.1,1,0.1),weights=NA,bg.corr="subtract",...)
object |
object of class “marrayRaw” or “marrayNorm” |
alpha |
vector of alpha parameters that are tested in the GCV procedure |
weights |
matrix of weights for local regression. Rows correspond to the spotted probe sequences, columns to arrays in the batch. These may be derived from the matrix of spot quality weights as defined for “marrayRaw” objects. |
bg.corr |
backcorrection method (for “marrayRaw” objects) : “none” or “subtract”(default). |
... |
Further arguments for |
The function oin is based on iterative local regression of logged fold changes
in respect to average logged spot intensities. It incorporates optimisation of the smoothing parameter alpha
that controls the neighbourhood size h of local fitting. The parameter alpha
specifies the fraction of points that are included in the neighbourhood and thus has a value between 0 and 1.
Larger alpha values lead to smoother fits.
If the normalisation should be based on set of genes assumed to be not differentially expressed (house-keeping genes), weights can be used for local regression. In this case, all weights should be set to zero except for the house-keeping genes for which weights are set to one. In order to achieve a reliable regression, it is important, however, that there is a sufficient number of house-keeping genes that are distributed over the whole expression range and spotted accross the whole array.
In contrast to OLIN and OSLIN, the OIN scheme does not correct for spatial dye bias. It can, therefore, be used if the assumption of random spotting does not hold.
Object of class “marrayNorm” with normalised logged ratios
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
maNorm, locfit,
gcv, olin
,lin, ino
# LOADING DATA data(sw) # OPTIMISED INTENSITY-DEPENDENT NORMALISATION norm.oin <- oin(sw) # MA-PLOT OF NORMALISATION RESULTS OF FIRST ARRAY plot(maA(norm.oin)[,1],maM(norm.oin)[,1],main="OIN") # CORRESPONDING MXY-PLOT mxy.plot(maM(norm.oin)[,1],Ngc=maNgc(norm.oin),Ngr=maNgr(norm.oin), Nsc=maNsc(norm.oin),Nsr=maNsr(norm.oin),main="OIN") ## LOADING DATA data(sw) # OPTIMISED INTENSITY-DEPENDENT NORMALISATION norm.oin <- oin(sw) # MA-PLOT OF NORMALISATION RESULTS OF FIRST ARRAY plot(maA(norm.oin)[,1],maM(norm.oin)[,1],main="OIN") # CORRESPONDING MXY-PLOT mxy.plot(maM(norm.oin)[,1],Ngc=maNgc(norm.oin),Ngr=maNgr(norm.oin), Nsc=maNsc(norm.oin),Nsr=maNsr(norm.oin),main="OIN") #
This functions performs optimised local intensity-dependent normalisation (OLIN) and optimised scaled intensity-dependent normalisation (OSLIN).
olin(object,X=NA,Y=NA,alpha=seq(0.1,1,0.1),iter=3, scale=c(0.05,0.1,0.5,1,2,10,20),OSLIN=FALSE,weights=NA, genepix=FALSE,bg.corr="subtract",...)olin(object,X=NA,Y=NA,alpha=seq(0.1,1,0.1),iter=3, scale=c(0.05,0.1,0.5,1,2,10,20),OSLIN=FALSE,weights=NA, genepix=FALSE,bg.corr="subtract",...)
object |
object of class “marrayRaw” or “marrayNorm” corresponding to a single array or a batch of arrays. |
X |
matrix with x-coordinates of spots of the arrays in |
Y |
matrix with y-coordinates of spots. Each column includes the y-coodinates for the spots of one array.If Y=NA, rows on array are used as proxies for the location in y-direction |
alpha |
vector of alpha parameters that are tested in the GCV procedure |
iter |
number of iterations in the OLIN procedure |
scale |
vector of scale parameters that are tested in a GCV procedure for spatial regression. This define the amount of smoothing in X-direction with respect to smoothing in Y-direction. |
OSLIN |
If OSLIN=TRUE, subsequent scaling of the range of M accross the array is performed. |
weights |
matrix of (non-negative) weights for local regression (see |
genepix |
If |
bg.corr |
backcorrection method (for “marrayRaw” objects) : “none”, “subtract”, “half”, “minimum”, “movingmin”, “edwards” or “normexp”. |
... |
Further arguments for |
OLIN and OSLIN are based on iterative local regression and incorporate optimisation of model parameters.
Local regression is performed using LOCFIT, which requires the user to choose a specific smoothing parameter alpha
that controls the neighbourhood size h of local fitting. The parameter alpha
specifies the fraction of points that are included in the neighbourhood and thus has a value between 0 and 1.
Larger alpha values lead to smoother fits.
Additionally, the setting of scale parameters controls for distinct amount of smoothing in
Y-direction compared to smoothing in X-direction.
The parameter scale can be of arbitrary value.
The choice of model parameters alpha and scale for local regression is crucial for the efficiency and
quality of normalization. To optimize the model parameters, a general cross-validation procedure (GCV) is applied.
The arguments alpha and scale define the parameters values
which are tested in the GCV. OSLIN comprises the OLIN procedure with a
subsequent optimized scaling of the range of logged intensity ratios
across the spatial dimensions of the array. Details concerning the
background correction methods can be found in the help page for backgroundCorrect2.
Detailed information about OLIN and OSLIN can be found in the package documentation and in the reference stated below.
The weights argument specifies the influence of the single spots on the local regression. To exclude
spots being used for the local regression (such as control spots), set their corresponding weight to zero.
Note that OLIN and OSLIN
are based on the assumptions that most genes are not differentially expressed (or up- and down-regulation
is balanced) and that genes are randomly spotted across the array. If these assumptions are not valid, local
regression can lead to an underestimation of differential expression. OSLIN is especially sensitive to violations of these assumptions. However, this
sensitivity can be decreased if the minimal alpha-value is increased. Minimal alpha defines the
smallest scale used for local regression. Increasing alpha can reduce the influence of localised
artifacts as a larger fraction of data points is included. Alternative normalisation functions such
as oin, lin and ino might also be used for a more conservative fit.
If the normalisation should be based on set of genes assumed to be not differentially expressed (house-keeping genes), weights can be used for local regression. In this case, all weights are set to zero except for the house-keeping genes for which weights are set to one. In order to achieve a reliable regression, it is important, however, that there is a sufficient number of house-keeping genes that are distributed over the whole expression range and spotted accross the whole array.
It is also important to note that OLIN/OSLIN is fairly efficient in removing intensity- and spatial-dependent dye bias, so that normalised data will look quite “good” after normalisation independently of the true underlying data quality. Normalisation by local regression assumes smoothness of bias. Therefore, localised artifacts such as scratches, edge effects or bubbles should be avoided. Spots of these areas should be flagged (before normalisation is applied) to ensure data integrity. To stringently detect artifacts, the OLIN functions fdr.int, fdr.spatial, p.int and p.spatial can be used.
Object of class “marrayNorm” with normalised logged ratios
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
M.Futschik and T.Crompton (2004) Model selection and efficiency testing for normalization of cDNA microarray data, Genome Biology, 5:R60
M.Futschik and T.Crompton (2005) OLIN: Optimized normalization, visualization and quality testing for two-channel microarray data, Bioinformatics, 21(8):1724-6
OLIN web-page: http://itb.biologie.hu-berlin.de/~futschik/software/R/OLIN
# LOADING DATA data(sw) data(sw.xy) # OPTIMISED LOCAL INTENSITY-DEPENDENT NORMALISATION OF FIRST ARRAY norm.olin <- olin(sw[,1],X=sw.xy$X[,1],Y=sw.xy$Y[,1]) # MA-PLOT OF NORMALISATION RESULTS OF FIRST ARRAY plot(maA(norm.olin),maM(norm.olin),main="OLIN") # CORRESPONDING MXY-PLOT mxy.plot(maM(norm.olin)[,1],Ngc=maNgc(norm.olin),Ngr=maNgr(norm.olin), Nsc=maNsc(norm.olin),Nsr=maNsr(norm.olin),main="OLIN") # OPTIMISED SCALED LOCAL INTENSITY-DEPENDENT NORMALISATION norm.oslin <- olin(sw[,1],X=sw.xy$X[,1],Y=sw.xy$Y[,1],OSLIN=TRUE) # MA-PLOT plot(maA(norm.oslin),maM(norm.oslin),main="OSLIN") # MXY-PLOT mxy.plot(maM(norm.oslin)[,1],Ngc=maNgc(norm.oslin),Ngr=maNgr(norm.oslin), Nsc=maNsc(norm.oslin),Nsr=maNsr(norm.oslin),main="OSLIN")# LOADING DATA data(sw) data(sw.xy) # OPTIMISED LOCAL INTENSITY-DEPENDENT NORMALISATION OF FIRST ARRAY norm.olin <- olin(sw[,1],X=sw.xy$X[,1],Y=sw.xy$Y[,1]) # MA-PLOT OF NORMALISATION RESULTS OF FIRST ARRAY plot(maA(norm.olin),maM(norm.olin),main="OLIN") # CORRESPONDING MXY-PLOT mxy.plot(maM(norm.olin)[,1],Ngc=maNgc(norm.olin),Ngr=maNgr(norm.olin), Nsc=maNsc(norm.olin),Nsr=maNsr(norm.olin),main="OLIN") # OPTIMISED SCALED LOCAL INTENSITY-DEPENDENT NORMALISATION norm.oslin <- olin(sw[,1],X=sw.xy$X[,1],Y=sw.xy$Y[,1],OSLIN=TRUE) # MA-PLOT plot(maA(norm.oslin),maM(norm.oslin),main="OSLIN") # MXY-PLOT mxy.plot(maM(norm.oslin)[,1],Ngc=maNgc(norm.oslin),Ngr=maNgr(norm.oslin), Nsc=maNsc(norm.oslin),Nsr=maNsr(norm.oslin),main="OSLIN")
This function assesses the significance of intensity-dependent bias. This is achieved by comparing the observed average values of logged fold-changes within an intensity neighbourhood with an empirical distribution generated by permutation tests. The significance is given by (adjusted) p-values.
p.int(A,M,delta=50,N=-1,av="median",p.adjust.method="none")p.int(A,M,delta=50,N=-1,av="median",p.adjust.method="none")
A |
vector of average logged spot intensity |
M |
vector of logged fold changes |
delta |
integer determining the size of the neighbourhood ( |
N |
number of random samples (of size |
av |
averaging of |
p.adjust.method |
method for adjusting p-values due to multiple testing regime. The available
methods are “none”, “bonferroni”, “holm”, “hochberg”,
“hommel” and “fdr”. See also |
The function p.int assesses the significance of intensity-dependent bias using a permutation test.
The null hypothesis states the independence of A and M. To test if M depends on A,
spots are ordered with respect to A. This defines a neighbourhood of spots with similar A for each spot.
Next, the test statistic is the median or mean of M
within
a spot's intensity neighbourhood of chosen size (2 *delta+1). The empirical distribution of the
this statistic is then generated based on N random samples (with replacement).
(Note that sampling without replacement is used for fdr.int. Also note, that different meaning of argument N
in p.int and fdr.int. The argument N in p.int is the number fo independent samples (of size 2 *delta+1)
derived from the original distribution. The argument N in fdr.int states how many times the original distribution
is randomised and the permutated distribution is used for generating the empirical distribution.)
Comparing this empirical distribution of
with the observed distribution of ,
the independence of M and A
is assessed. If M is independent of A, the empirical distribution
of can be expected to be symmetrically
distributed around its mean value. To assess the significance of observing positive deviations of
the p-values are used. It indicates the expected proportion of neighbourhoods with larger
than the actual one based on the empirical distribution of
. The minimal p-value is set to 1/N.
Correspondingly, the significance
of observing negative deviations of can be determined.
Since this assessment of significance involves multiple testing, an adjustment of the p-values might be advisable.
A list of vector containing the p-values for positive (Pp) and negative (Pn) deviations of
of the spot's neighbourhood is produced. Values corresponding to spots
within an interval of delta at the lower or upper end of the A-scale are set to NA.
The same functionality but with our input and output formats is offered by p.int2
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
p.int2,fdr.int, sigint.plot, p.adjust
# To run these examples, "un-comment" them! # # LOADING DATA NOT-NORMALISED # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # P <- p.int(maA(sw)[,1],maM(sw)[,1],delta=50,N=10000,av="median",p.adjust.method="none") # VISUALISATION OF RESULTS # sigint.plot(maA(sw)[,1],maM(sw)[,1],Sp=P$Pp,Sn=P$Pn,c(-5,-5)) # LOADING NORMALISED DATA # data(sw.olin) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # P <- p.int(maA(sw.olin)[,1],maM(sw.olin)[,1],delta=50,N=10000,av="median",p.adjust.method="none") # VISUALISATION OF RESULTS # sigint.plot(maA(sw.olin)[,1],maM(sw.olin)[,1],Sp=P$Pp,Sn=P$Pn,c(-5,-5))# To run these examples, "un-comment" them! # # LOADING DATA NOT-NORMALISED # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # P <- p.int(maA(sw)[,1],maM(sw)[,1],delta=50,N=10000,av="median",p.adjust.method="none") # VISUALISATION OF RESULTS # sigint.plot(maA(sw)[,1],maM(sw)[,1],Sp=P$Pp,Sn=P$Pn,c(-5,-5)) # LOADING NORMALISED DATA # data(sw.olin) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # P <- p.int(maA(sw.olin)[,1],maM(sw.olin)[,1],delta=50,N=10000,av="median",p.adjust.method="none") # VISUALISATION OF RESULTS # sigint.plot(maA(sw.olin)[,1],maM(sw.olin)[,1],Sp=P$Pp,Sn=P$Pn,c(-5,-5))
This function assesses the significance of intensity-dependent bias. This is achieved by comparing the observed average values of logged fold-changes within an intensity neighbourhood with an empirical distribution generated by permutation tests. The significance is given by (adjusted) p-values.
p.int2(object,delta=50,N=-1,av="median",p.adjust.method="none")p.int2(object,delta=50,N=-1,av="median",p.adjust.method="none")
object |
object of class marrayRaw or marrayNorm |
delta |
integer determining the size of the neighbourhood ( |
N |
number of random samples (of size |
av |
averaging of |
p.adjust.method |
method for adjusting p-values due to multiple testing regime. The available
methods are “none”, “bonferroni”, “holm”, “hochberg”,
“hommel” and “fdr”. See also |
This function p.int2 is basically the same as p.int except for
differences in their in- and output format. For the details about the functionality, see p.int.
This function will be merged with p.int in future versions.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
p.int,fdr.int2, sigint.plot2, p.adjust
# To run these examples, "un-comment" them! # # LOADING DATA NOT-NORMALISED # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # P <- p.int2(sw,delta=50,N=10000,av="median",p.adjust.method="none") # VISUALISATION OF RESULTS # sigint.plot2(sw[,1],Sp=P$Pp[[1]],Sn=P$Pn[[1]],c(-5,-5)) # array 1 # sigint.plot2(sw[,3],Sp=P$Pp[[3]],Sn=P$Pn[[3]],c(-5,-5)) # array 3# To run these examples, "un-comment" them! # # LOADING DATA NOT-NORMALISED # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # P <- p.int2(sw,delta=50,N=10000,av="median",p.adjust.method="none") # VISUALISATION OF RESULTS # sigint.plot2(sw[,1],Sp=P$Pp[[1]],Sn=P$Pn[[1]],c(-5,-5)) # array 1 # sigint.plot2(sw[,3],Sp=P$Pp[[3]],Sn=P$Pn[[3]],c(-5,-5)) # array 3
This function assesses the significance of spatial bias. This is achieved by comparing the observed average values of logged fold-changes within a spot's spatial neighbourhood with an empirical distribution generated by permutation tests. The significance is given by (adjusted) p-values derived in one-sided permutation test.
p.spatial(X,delta=2,N=-1,av="median",p.adjust.method="none")p.spatial(X,delta=2,N=-1,av="median",p.adjust.method="none")
X |
matrix of logged fold changes |
delta |
integer determining the size of spot neighbourhoods
( |
N |
number of samples for generation of empirical background distribution |
av |
averaging of |
p.adjust.method |
method for adjusting p-values due to multiple testing regime. The available
methods are “none”, “bonferroni”, “holm”, “hochberg”,
“hommel” and “fdr”. See also |
The function p.spatial assesses the significance of spatial bias using an one-sided random
permutation test.
The null hypothesis states random spotting i.e. the independence of log ratio M
and spot location. First, a neighbourhood of a spot is defined by a two dimensional square window
of chosen size ((2*delta+1)x(2*delta+1)). Next, a test statistic is defined by calculating
the median or mean of M for N random samples
of size ((2*delta+1)x(2*delta+1)). Note that this scheme defines a sampling with replacement
procedure whereas sampling without replacement is used for fdr.spatial.
Comparing the empirical distribution of
with the observed distribution of ,
the independence of M and spot location
can be assessed. If M is independent of spot's location,
the empirical distribution can be expected to be
distributed around its mean value. To assess the significance of observing positive deviations of
,
p-values are calculated using Fisher's method. The p-value equals the fraction of values in the empirical
distribution which are larger than the observed value . The minimal p-value is set to 1/N.
Correspondingly, the significance
of observing negative deviations of can be determined.
A list of vectors containing the p-values for positive (Pp)
and negative (Pn) deviations of
of the spot's neighbourhood is produced.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
# To run these examples, "un-comment" them! # # LOADING DATA # data(sw) # M <- v2m(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw), # Nsc=maNsc(sw),Nsr=maNsr(sw),main="MXY plot of SW-array 1") # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # P <- p.spatial(M,delta=2,N=10000,av="median") # sigxy.plot(P$Pp,P$Pn,color.lim=c(-5,5),main="FDR") # LOADING NORMALISED DATA # data(sw.olin) # M <- v2m(maM(sw.olin)[,1],Ngc=maNgc(sw.olin),Ngr=maNgr(sw.olin), # Nsc=maNsc(sw.olin),Nsr=maNsr(sw.olin),main="MXY plot of SW-array 1") # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # P <- p.spatial(M,delta=2,N=10000,av="median") # VISUALISATION OF RESULTS # sigxy.plot(P$Pp,P$Pn,color.lim=c(-5,5),main="FDR")# To run these examples, "un-comment" them! # # LOADING DATA # data(sw) # M <- v2m(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw), # Nsc=maNsc(sw),Nsr=maNsr(sw),main="MXY plot of SW-array 1") # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # P <- p.spatial(M,delta=2,N=10000,av="median") # sigxy.plot(P$Pp,P$Pn,color.lim=c(-5,5),main="FDR") # LOADING NORMALISED DATA # data(sw.olin) # M <- v2m(maM(sw.olin)[,1],Ngc=maNgc(sw.olin),Ngr=maNgr(sw.olin), # Nsc=maNsc(sw.olin),Nsr=maNsr(sw.olin),main="MXY plot of SW-array 1") # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # P <- p.spatial(M,delta=2,N=10000,av="median") # VISUALISATION OF RESULTS # sigxy.plot(P$Pp,P$Pn,color.lim=c(-5,5),main="FDR")
This function assesses the significance of spatial bias. This is achieved by comparing the observed average values of logged fold-changes within a spot's spatial neighbourhood with an empirical distribution generated by permutation tests. The significance is given by (adjusted) p-values derived in one-sided permutation test.
p.spatial2(object,delta=2,N=-1,av="median",p.adjust.method="none")p.spatial2(object,delta=2,N=-1,av="median",p.adjust.method="none")
object |
object of class marrayRaw or marrayNorm |
delta |
integer determining the size of spot neighbourhoods
( |
N |
number of samples for generation of empirical background distribution |
av |
averaging of |
p.adjust.method |
method for adjusting p-values due to multiple testing regime. The available
methods are “none”, “bonferroni”, “holm”, “hochberg”,
“hommel” and “fdr”. See also |
The function p.spatial2.Rd is basically the same as p.spatial,
but differs in its input and output formats. Details about the functionality can be found
at p.spatial.
A list of a two lists of vectors is produced containing the p-values for positive (Pp)
and negative (Pn) deviations of
of the spot's neighbourhood is produced (see example below).
This function will be fused with p.spatial in future versions using S4-style methods.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
fdr.int, sigxy.plot, p.adjust,p.spatial
# To run these examples, "un-comment" them! # # LOADING DATA # data(sw) # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # P <- p.spatial2(sw,delta=2,N=10000,av="median") # SIGNIFICANCE PLOTS OF ARRAY 1 # sigxy.plot2(sw[,1],P$Pp[[1]],P$Pn[[1]],color.lim=c(-5,5),main="P-value") # SIGNIFICANCE PLOTS OF ARRAY 3 # sigxy.plot2(sw[,3],P$Pp[[3]],P$Pn[[3]],color.lim=c(-5,5),main="P-value")# To run these examples, "un-comment" them! # # LOADING DATA # data(sw) # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # P <- p.spatial2(sw,delta=2,N=10000,av="median") # SIGNIFICANCE PLOTS OF ARRAY 1 # sigxy.plot2(sw[,1],P$Pp[[1]],P$Pn[[1]],color.lim=c(-5,5),main="P-value") # SIGNIFICANCE PLOTS OF ARRAY 3 # sigxy.plot2(sw[,3],P$Pp[[3]],P$Pn[[3]],color.lim=c(-5,5),main="P-value")
This function sets data to NA if the corresponding spots have significantly biased neighbourhoods on the intensity scale or on the spatial dimensions of the array.
sig.mask(object,Sp,Sn,thrp,thrn)sig.mask(object,Sp,Sn,thrp,thrn)
object |
object of class |
Sp |
list of vectors of false discovery rate or p-values for positive deviation of |
Sn |
list vector of false discovery rate or p-values for negative deviation of |
thrp |
vector of thresholds for significance of positive deviation ( |
thrn |
vector of thresholds for significance of negative deviation ( |
This function can be used for the masking of data that has been decided to be unrelaible
after the application of significance test for intenstiy- and location dependent dye bias (e.g. p.int2, fdr.int2,
p.spatial2, fdr.spatial2).
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
sigint.plot, fdr.int, p.int, sigxy.plot,
fdr.spatial, p.spatial
# To run these commands, delete comment sign (#) ! # # LOADING DATA # data(sw) # # MASKING REGIONS WITH SPATIAL DYE BIAS # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this example, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.spatial2(sw,delta=2,N=10,av="median",edgeNA=FALSE) # # VISUALISATION # sigxy.plot2(sw[,1],FDR$FDRp[[1]],FDR$FDRn[[1]],color.lim=c(-5,5),main="FDR") # # MASKING SIGNIFICANT NEIGHBOURHOODS # thresp <- c(0.01,0.01,0.01,0.01) # thresn <- c(0.01,0.01,0.01,0.01) # sw.masked <- sig.mask(sw,Sp=FDR$FDRp,Sn=FDR$FDRn,thrp=thresp,thrn=thresn) # mxy.plot(sw.masked[,4]) # plot masked data for array 4# To run these commands, delete comment sign (#) ! # # LOADING DATA # data(sw) # # MASKING REGIONS WITH SPATIAL DYE BIAS # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this example, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.spatial2(sw,delta=2,N=10,av="median",edgeNA=FALSE) # # VISUALISATION # sigxy.plot2(sw[,1],FDR$FDRp[[1]],FDR$FDRn[[1]],color.lim=c(-5,5),main="FDR") # # MASKING SIGNIFICANT NEIGHBOURHOODS # thresp <- c(0.01,0.01,0.01,0.01) # thresn <- c(0.01,0.01,0.01,0.01) # sw.masked <- sig.mask(sw,Sp=FDR$FDRp,Sn=FDR$FDRn,thrp=thresp,thrn=thresn) # mxy.plot(sw.masked[,4]) # plot masked data for array 4
This function visualises the significance of intensity-dependent bias.
sigint.plot(A,M,Sp,Sn,ylim=c(-3,-3),...)sigint.plot(A,M,Sp,Sn,ylim=c(-3,-3),...)
A |
vector of average logged spot intensity |
M |
vector of logged fold changes |
Sp |
vector of false discovery rate or p-values for positive deviation of |
Sn |
vector of false discovery rate or p-values for negative deviation of |
ylim |
vector of minimal log10(fdr) or log10(p-value) to be visualised corresponding to |
... |
Further optional graphical parameter for the |
The function sigint.plot produces a MA-plot of the significance (Sp,Sn)
generated by fdr.int or p.int. The abscissa (x-axis) is shows by the average logged spot intensity
A=0.5*(log(Cy3)+log(Cy5)); the ordinate axis (y-axis) shows the log10(FDR) or log10(p) given by
FDRp or Pn and FDRn or Pn.
The significance for positive of
spot intensity neighbourhoods are presented by red colour; the significance for negative
of
spot intensity neighbourhoods are presented by green colour. The ordinate axis (y-axis) give the
log10-transformed FDR or p-values.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
# To run these examples, "un-comment" them! # # LOADING DATA NOT-NORMALISED # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # This can take a while! For testing, you may choose a smaller N. # FDR <- fdr.int(maA(sw)[,1],maM(sw)[,1],delta=50,N=100,av="median") # VISUALISATION OF RESULTS # sigint.plot(maA(sw)[,1],maM(sw)[,1],FDR$FDRp,FDR$FDRn,c(-5,-5)) # data(sw.olin) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # F <- fdr.int(maA(sw.olin)[,1],maM(sw.olin)[,1],delta=50,N=100,av="median") # VISUALISATION OF RESULTS # sigint.plot(maA(sw.olin)[,1],maM(sw.olin)[,1],FDR$FDRp,FDR$FDRn,c(-5,-5))# To run these examples, "un-comment" them! # # LOADING DATA NOT-NORMALISED # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # This can take a while! For testing, you may choose a smaller N. # FDR <- fdr.int(maA(sw)[,1],maM(sw)[,1],delta=50,N=100,av="median") # VISUALISATION OF RESULTS # sigint.plot(maA(sw)[,1],maM(sw)[,1],FDR$FDRp,FDR$FDRn,c(-5,-5)) # data(sw.olin) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # F <- fdr.int(maA(sw.olin)[,1],maM(sw.olin)[,1],delta=50,N=100,av="median") # VISUALISATION OF RESULTS # sigint.plot(maA(sw.olin)[,1],maM(sw.olin)[,1],FDR$FDRp,FDR$FDRn,c(-5,-5))
This function produce visualises the significance of intensity-dependent bias.
sigint.plot2(object,Sp,Sn,ylim=c(-3,-3),...)sigint.plot2(object,Sp,Sn,ylim=c(-3,-3),...)
object |
object of class marrayRaw or marrayNorm |
Sp |
vector of false discovery rate or p-values for positive deviation of |
Sn |
vector of false discovery rate or p-values for negative deviation of |
ylim |
vector of minimal log10(fdr) or log10(p-value) to be visualised corresponding to |
... |
Further optional graphical parameter for the |
The function sigint.plot2 only differs from sigint.plot in its input arguments.
The functionality is the same. For details, see sigint.plot.
This function will be merged with sigint.plot in future versions.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
# To run these examples, delete the comment signs (#) in front of the commands. # # LOADING DATA NOT-NORMALISED # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this example, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.int2(sw,delta=50,N=10,av="median") # VISUALISATION OF RESULTS # sigint.plot2(sw[,1],FDR$FDRp[[1]],FDR$FDRn[[1]],c(-5,-5)) # array 1 # sigint.plot2(sw[,4],FDR$FDRp[[4]],FDR$FDRn[[4]],c(-5,-5)) # array 4# To run these examples, delete the comment signs (#) in front of the commands. # # LOADING DATA NOT-NORMALISED # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this example, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.int2(sw,delta=50,N=10,av="median") # VISUALISATION OF RESULTS # sigint.plot2(sw[,1],FDR$FDRp[[1]],FDR$FDRn[[1]],c(-5,-5)) # array 1 # sigint.plot2(sw[,4],FDR$FDRp[[4]],FDR$FDRn[[4]],c(-5,-5)) # array 4
This function produces a 2D-plot visualizing the significance of spatial bias.
sigxy.plot(Sp,Sn,color.lim=c(-3,3),...)sigxy.plot(Sp,Sn,color.lim=c(-3,3),...)
Sp |
matrix of false discovery rates or p-values for positive deviation of
|
Sn |
matrix of false discovery rate or p-values for negative deviation of |
color.lim |
limits of color range for plotting vector corresponding to log10( |
... |
Further optional graphical parameter for the |
The function sigxy.plot produces a 2d-plot presenting the significance (pS,nS)
generated by fdrint or p.spatial.
The significance Sp for positive of
spatial spot neighbourhoods are presented by red colour; the significance(Sn) for negative
of
spatial spot neighbourhoods are presented by green colour.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
colorbar.sig, fdr.spatial, p.spatial, image,
p.spatial
# To run these examples, "un-comment" them! # # LOADING DATA # data(sw) # # M <- v2m(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw), # Nsc=maNsc(sw),Nsr=maNsr(sw),main="MXY plot of SW-array 1") # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # This can take a while! For testing, you may choose a smaller N. # FDR <- fdr.spatial(M,delta=2,N=100,av="median",edgeNA=TRUE) # sigxy.plot(FDR$FDRp,FDR$FDRn,color.lim=c(-5,5),main="FDR") # # LOADING NORMALISED DATA # data(sw.olin) # M <- v2m(maM(sw.olin)[,1],Ngc=maNgc(sw.olin),Ngr=maNgr(sw.olin), # Nsc=maNsc(sw.olin),Nsr=maNsr(sw.olin),main="MXY plot of SW-array 1") # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # FDR <- fdr.spatial(M,delta=2,N=100,av="median",edgeNA=TRUE) # VISUALISATION OF RESULTS # sigxy.plot(FDR$FDRp,FDR$FDRn,color.lim=c(-5,5),main="FDR") # # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # P <- p.spatial(M,delta=2,N=-1,av="median",p.adjust.method="holm") # VISUALISATION OF RESULTS # sigxy.plot(P$Pp,P$Pn,color.lim=c(-5,5),main="FDR")# To run these examples, "un-comment" them! # # LOADING DATA # data(sw) # # M <- v2m(maM(sw)[,1],Ngc=maNgc(sw),Ngr=maNgr(sw), # Nsc=maNsc(sw),Nsr=maNsr(sw),main="MXY plot of SW-array 1") # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # This can take a while! For testing, you may choose a smaller N. # FDR <- fdr.spatial(M,delta=2,N=100,av="median",edgeNA=TRUE) # sigxy.plot(FDR$FDRp,FDR$FDRn,color.lim=c(-5,5),main="FDR") # # LOADING NORMALISED DATA # data(sw.olin) # M <- v2m(maM(sw.olin)[,1],Ngc=maNgc(sw.olin),Ngr=maNgr(sw.olin), # Nsc=maNsc(sw.olin),Nsr=maNsr(sw.olin),main="MXY plot of SW-array 1") # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # FDR <- fdr.spatial(M,delta=2,N=100,av="median",edgeNA=TRUE) # VISUALISATION OF RESULTS # sigxy.plot(FDR$FDRp,FDR$FDRn,color.lim=c(-5,5),main="FDR") # # # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # P <- p.spatial(M,delta=2,N=-1,av="median",p.adjust.method="holm") # VISUALISATION OF RESULTS # sigxy.plot(P$Pp,P$Pn,color.lim=c(-5,5),main="FDR")
This function produces a 2D-plot visualizing the significance of spatial bias.
sigxy.plot2(object,Sp,Sn,color.lim=c(-3,3),...)sigxy.plot2(object,Sp,Sn,color.lim=c(-3,3),...)
object |
object of class marrayRaw or marrayNorm |
Sp |
vector of false discovery rates or p-values for positive deviation of
|
Sn |
vector of false discovery rate or p-values for negative deviation of |
color.lim |
limits of color range for plotting vector corresponding to log10( |
... |
Further optional graphical parameter for the |
The function sigxy.plot2 differs from sigxy.plot in its input arguments.
The functionality is the same. For details, see sigxy.plot.
This function will be merged with sigxy.plot in future versions.
Matthias E. Futschik (http://itb.biologie.hu-berlin.de/~futschik)
colorbar.sig, sigxy.plot,
sigxy.plot,fdr.spatial2, p.spatial2, image
# To run these examples, "un-comment" them! # # LOADING DATA # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.spatial2(sw,delta=2,N=10,av="median",edgeNA=TRUE) # # SIGNIFICANCE PLOTS OF ARRAY 1 # sigxy.plot2(sw[,1],FDR$FDRp[[1]],FDR$FDRn[[1]],color.lim=c(-5,5),main="FDR") # SIGNIFICANCE PLOTS OF ARRAY 3 # sigxy.plot2(sw[,3],FDR$FDRp[[3]],FDR$FDRn[[3]],color.lim=c(-5,5),main="FDR") ## To run these examples, "un-comment" them! # # LOADING DATA # data(sw) # CALCULATION OF SIGNIFICANCE OF SPOT NEIGHBOURHOODS # For this illustration, N was chosen rather small. For "real" analysis, it should be larger. # FDR <- fdr.spatial2(sw,delta=2,N=10,av="median",edgeNA=TRUE) # # SIGNIFICANCE PLOTS OF ARRAY 1 # sigxy.plot2(sw[,1],FDR$FDRp[[1]],FDR$FDRn[[1]],color.lim=c(-5,5),main="FDR") # SIGNIFICANCE PLOTS OF ARRAY 3 # sigxy.plot2(sw[,3],FDR$FDRp[[3]],FDR$FDRn[[3]],color.lim=c(-5,5),main="FDR") #
Gene expression in two cancer cell lines, SW480 and SW620, is compared. The SW480 cell line was derived from a colon tumour of a 50-year old male patient. The second cell line (SW620) originated from a lymph node metastasis of the same patient. Sharing the same genetic background, these cell lines serve as an model of cancer progression.
Target cDNA from SW480 was labelled with Cy3 whereas cDNA from SW620 was labelled with Cy5 using the amino-allyl labelling method. Both cDNA pools were co-hybridised on glass slides with 8448 spots. The spots consisted of 3986 distinct sequence-verified human cDNA clones (Research Genetics, release GF211) printed in duplicates, 84 spots from non-human cDNA clones and a further 154 control spots. Spots were printed by 4x4 pins. The experiment consisted of four replicated arrays derived from separate labelling reactions. Local background spot intensities were extracted by QuantArray software (version2.1). Analysis showed that replicated spots were highly correlated (average Pearson correlation: 0.94). Since this may interfere with the efficiency testing performed (and to reduce the size of the data set for illustration purpose), the replicated spots were not included here. Experimental details and further analysis can be found in Futschik et al. (2002).
data(sw)data(sw)
An object of class “marrayRaw”
The data was produced and provided by Sharon Pattison of the Cancer Genetics lab and Aaron Jeffs of the Otago Genomics facility of the University of Otago, Dunedin, New Zealand.
M. Futschik, A.Jeffs, S.Pattison, N.Kasabov, M.Sullivan, A.Merrie and A.Reeve, Gene expression profiling of metastatic and non-metastatic colorectal cancer cell-lines, Genome Letters, vol.1, No.1, pp. 26-34, 2002
The data set sw.olin is derived from data set sw
by optimised local intensity-dependent normalisation (OLIN).
data(sw.olin)data(sw.olin)
An object of class “marrayNorm”
The original data (sw) was produced and provided by S.Pattison of the Cancer Genetics lab and A. Jeffs of
the Otago Genomics facility of the University of Otago, Dunedin, New Zealand.
M. Futschik, A.Jeffs, S.Pattison, N.Kasabov, M.Sullivan, A.Merrie and A.Reeve, Gene expression profiling of metastatic and non-metastatic colorectal cancer cell-lines, Genome Letters, vol.1, No.1, pp. 26-34, 2002
The data set sw.xy contains the x- and y-coordinates of the spots
in the data set sw
data(sw.xy)data(sw.xy)
A list of two matrices
The original data (sw) was produced and provided by S.Pattison of the Cancer Genetics lab and A.Jeffs of
the Otago Genomics facility of the University of Otago, Dunedin, New Zealand.
M. Futschik, A.Jeffs, S.Pattison, N.Kasabov, M.Sullivan, A.Merrie and A.Reeve, Gene expression profiling of metastatic and non-metastatic colorectal cancer cell-lines, Genome Letters, vol.1, No.1, pp. 26-34, 2002
This functions converts a vector to a matrix based on a given spot layout. Optionally, it produces a 2D-plot.
v2m(V,Ngc,Ngr,Nsc,Nsr,visu=FALSE,color.lim=c(-1,1),xlab="Columns",ylab="Rows",...)v2m(V,Ngc,Ngr,Nsc,Nsr,visu=FALSE,color.lim=c(-1,1),xlab="Columns",ylab="Rows",...)
V |
vector of real values |
Ngc |
number of columns for the grid matrix |
Ngr |
number of rows for the grid matrix |
Nsc |
number of columns for the spot matrix |
Nsr |
number of rows for the spot matrix |
visu |
If FALSE, MXY plot is generated. |
color.lim |
Limits of color range for MXY plot |
xlab |
label of x -axis of MXY plot |
ylab |
label of y-axis of MXY plot |
... |
Further optional parameters for the |
The function v2m converts a vector V (as e.g. derived by maM(object)[,index]) to a matrix representing
the spatial distribution of the values of V across the array.
Note that this function assumes a specific mapping between the data points and the location of spot (i.e.
the same mapping rule that is used for marrayRaw/marrayNorm objects.) The validity of this mapping should be
carefully checked (see also the documentation of packet marray.)
The option for spatial visualisation
is rather restricted to logged fold-changes as the corresponding colour range is centred around zero and
follows the conventional colouring (green for negative, red for positive fold-changes). The MXY plot produced
by v2n does not include a colour bar. To have a colour included, you can use mxy.plot.
A 2D-matrix with (Ngc*Nsc) columns and (Ngr*Nsr) is produced. This matrix represents the spatial
distribution of the values of vector V given the print-layout.
Matthias E. Futschik, http://itb.biologie.hu-berlin.de/~futschik
# LOADING DATA NOT-NORMALISED data(sw.olin) # CONVERSION FROM VECTOR TO MATRIX M <- v2m(maM(sw.olin)[,1],Ngc=maNgc(sw.olin),Ngr=maNgr(sw.olin), Nsc=maNsc(sw.olin),Nsr=maNsr(sw.olin),visu=TRUE) # BACK-CONVERSION FROM MATRIX TO VECTOR V <- m2v(M,Ngc=maNgc(sw.olin),Ngr=maNgr(sw.olin), Nsc=maNsc(sw.olin),Nsr=maNsr(sw.olin),visu=TRUE)# LOADING DATA NOT-NORMALISED data(sw.olin) # CONVERSION FROM VECTOR TO MATRIX M <- v2m(maM(sw.olin)[,1],Ngc=maNgc(sw.olin),Ngr=maNgr(sw.olin), Nsc=maNsc(sw.olin),Nsr=maNsr(sw.olin),visu=TRUE) # BACK-CONVERSION FROM MATRIX TO VECTOR V <- m2v(M,Ngc=maNgc(sw.olin),Ngr=maNgr(sw.olin), Nsc=maNsc(sw.olin),Nsr=maNsr(sw.olin),visu=TRUE)