anovaspatial: One-factorial ANOVA assessing spatial bias
In OLIN: Optimized local intensity-dependent normalisation of two-color microarrays

Description Usage Arguments Details Value Author(s) See Also Examples

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.

1	anovaspatial(obj,index,xN=5,yN=5,visu=FALSE)

`obj`	object of class “marrayRaw” or “marrayNorm”
`index`	index of array (within `obj`) to be tested
`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 M ~ (A - 1) (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 R-square 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