发信人: leohawk (leohawk), 信区: Biology
标 题: Re: microarray analysis for help!
发信站: Unknown Space - 未名空间 (Fri Nov 19 21:44:19 2004), 转信

well, I just read all the replies, seems no one asked you how
many chips you did and what is your exact defination of
down or up regulation, algorithm used, etc.

clustering, in most cases, does not make biological sense at
all. What you can do is to select the list of genes with good
P value and use Affymetrix's website for GO analysis, it is hard
to save results but kinda give you a tree like structure based
on your genes GO identity.

for probe level analysis: Dchip, RMA, MAS5.

For higher level analysis: Dchip, Genespring, ArrayTools, SAM,
genecluster2, etc, there are so many of them. I personally think
Genespring is no good for finding gene expression changes, but a
perfect tool for visualization or integration of multiple results.

ArrayTools implement a class comparsion that uses permutation test,
I find it the most useful if you have a lot of samples.

for things you wanna do, check this:

http://www.medinfopoli.polimi.it/GFINDer/

good luck.

【 在 lanceandfei (La Jolla的云彩) 的大作中提到: 】
: I got some microarray datas, in culture neurons.
: stimulation almost changes gene expression pattern totally.
: so upregulate or downregulate genes are at least 2000!!!
: boss pushed me to cluster them with function, receptor, transcription
: factors....
: I talked to many people, but no one could give me some straight ways.
: anybody know sth about that?
: Thanks so much in advance!!!!!!!!!!


--
※ 来源:．Unknown Space - 未名空间 mitbbs.com．[FROM: 68.38.]