[R--gR] Hello!

[zhihua li]

Hi,Dr.Zhou

Nice to receive your letter. If A activates B, it is conceivable that their 
expression profiles tend to be similar. But this seems to arise from a 
common sense or thumb rule. I've never seen a paper studying the details of 
"similarity" of expression profiles. Instead different people use their own 
defination of similatiry and develop different kinds of algorithms, like 
euclean distance in clustering, or conditinal propability in BNT.
As for the golden pathway to be tested usint the new algorithms, I think 
the starting point is other people's published datasets. There are numerous 
papers emerging in bioinforamtics/computational biology journals describing 
different algorithms every month. Almost each of them uses some datasets 
and presents their re-constructed networks. I think you can run your 
algorithm with their datasets and compare your results with theirs and 
known biological facts. In fact, I don't think there exists a "golden 
pathway" which is the touchstone of modeling genetic networks. 

Hope this may help.

Li



>From: "Dr. Zonglin Zhou" <z6zhou at bioinformatics.uwaterloo.ca>
>To: "'lzhtom'" <lzhtom at hotmail.com>
>Subject: Hello!
>Date: Sat, 5 Nov 2005 17:55:25 -0500
>