[BioC] tilingArray problems

Anjan Purkayastha purkayas at wi.mit.edu
Fri Nov 7 18:41:50 CET 2008

Ran tilingArray package on my dataset. But instead of the well-separated 
signal and noise intensities a la Huber et al I get a get a noisy data 
transformation (see  normalizedbytilingArray). As a comparison I have 
plotted the Non-normalized signal intensities, Reference signal 
intensities and the Non-normalized intensity-divided-by-Reference 
intensity in the second plot (tilingArray_Normalization_1). It looks 
like the strong DNA hyb reference signal may be masking the differences 
in the signal intensities between transcribed and untranscribed regions.
Anyone else having similar problems.
Brief description of the array: it consists of non-overlapping 60mers 
tiled along the vaccinia virus  genome sequence (ca 190kb)- there are 
about 6200 probes in all. We used 3micrograms of DNA for the reference 
array hybs and 1.2 micrograms of RNA for the expression hybs.
Normalization was done with the weakest 5% of probes dropped.

Thanks in advance.

anjan purkayastha, phd
bioinformatics analyst
whitehead institute for biomedical research
nine cambridge center
cambridge, ma 02142

purkayas [at] wi [dot] mit [dot] edu  

More information about the Bioconductor mailing list