Question: Correlation among genes in micro array data
0
gravatar for izsyed16
2.8 years ago by
izsyed1610
izsyed1610 wrote:

Hi all, I have a basic question regarding correlation calculation for gene micro array data. I have micro array data with gene names in rows and columns containing gene expression value. Now there are 2000 genes and 8 samples with 4 samples for healthy data and 4 samples for disease data. For correlation calculation among two genes should I have to consider only disease related 4 samples or should I consider all 8 samples at once and calculate correlation among genes. I want to use Matlab for correlation calculation but no idea which sample should I pick for this purpose. Also If there are more than two categories like healthy data 1, healthy data 2, disease data 1, disease data 1, then how should I continue with it. Thanks.

ADD COMMENTlink modified 2.8 years ago by Jean-Karim Heriche23k • written 2.8 years ago by izsyed1610

It depends on the question you want to address with this. Pearson's correlation tells you how strong the linear relationship is between the expression levels of two genes. In which context are you interested in this ? Do you want to know if two genes are correlated in general or in healthy people only or in disease 1 only ? The computation of any quantity should be motivated by a biological question. Find the question you want to address then compute the relevant answer.

ADD REPLYlink written 2.8 years ago by Jean-Karim Heriche23k

I want to make a network based on correlation among genes. If i want to find disease module in this network then should I consider only diseased samples. And if I want to know that how two genes are correlated in general then I will consider both diseased and healthy sample?

ADD REPLYlink written 2.8 years ago by izsyed1610

Yes, if you only care about correlation in general, you take all data.

ADD REPLYlink written 2.8 years ago by Jean-Karim Heriche23k
0
gravatar for Jean-Karim Heriche
2.8 years ago by
EMBL Heidelberg, Germany
Jean-Karim Heriche23k wrote:

To find disease-specific modules, one approach would be to build two networks, one for healthy and one for disease then try and find the modules that differ between them. For this, you could use a tensor factorization approach based on the CP decomposition. For more details, see my tutorial on the topic.

ADD COMMENTlink written 2.8 years ago by Jean-Karim Heriche23k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 834 users visited in the last hour