Narrowing to an actual binding site from ChIP-Chip data - How do you really read this stuff? (Sep/08/2008 )

Hello all - here's my question for you:

My ChIP-Chip data has given me a graph of bound regions for my protein of interest (in E. coli), but as you zoom in closer and closer some of these regions are relatively large - like 5 kb or even 10 kb of what is indicated as 'bound' by the protein. So how in the heck can I narrow these regions down to a binding sequence motif? Or is this microarray data incorrect? What is the normal 'width' of peak data that any of you normally get? I've tried several motif-finding algorithms but with input sequences so big, there are numerous possibilities produced.

Any suggestions?

Platform/Software:
Nimblegen E Coli K-12 Whole-genome tiled microarray
NimbleScan/SignalMap software

Thank you.

I have similar issue Nimblegen is not very helpful especially there technical support has very limited scientific knowledge . Even after full service nimblegen does not provide and access to raw data or nimble scan software. My question is:
I have data of chip on chip with peak ratio in excel files, along with chromosome #, start and stop of probe peak. Now how to further analyze it I believe one has to apply some algorithms (statistical analysis) to find out significant sites. Any one can please direct to me some available software or someone has experience in analyzing nimblegen chip on chip data. I talked with agilent but there software cannot take nimblgen's data as such.

Thanks

Is it available for download or you need to purcahse it? I googled the serach http://www.bioconductor.org/docs/install/
but there are various versions, which on eto use. Secondly what kind of files we might need.
Thanks