different kind of comparisons

Here we discuss using the GenEx Data Manager

Moderator: MultiD Support

different kind of comparisons

Postby MultiD Support » Fri Jul 04, 2008 9:19 am

Hello

I need one advice for mi data management.
I am working with three varieties of one plant, and I am doing two kinds of comparisons: 1. Comparison of expression of different genes between two tissues (leaves and roots) in each plant variety (paired experiment). 2. Comparison of expression of the same genes but among three varieties of the plant. In this way, I am doing comparison of the expression of theses gene among the leaves or roots of the three varieties (non-paired experiment)

I am introducing all my data in Genex, and for the first comparison, I am getting relative quantities with the average of expression, using the data of expression from leaves and roots together.

However in the second kind of comparison (among plant varieties), I am not sure, if I could use the same relative quantities that I got in the first comparison. In this case I want compare the values of expression of leaves in three varieties independent from the values of the roots (it is another comparison). What must I to do?

Finally, what should I do for a multivariable analysis (4 genes)? In this way, could I use the relative quantities obtained in the first comparison or must I do a multivariable analysis in each tissue (independently) using the data from the three varieties together?

Thank you very much for you helps.

Morgan Echeverry
MultiD Support
 
Posts: 170
Joined: Tue Jul 01, 2008 11:49 am

Re: different kind of comparisons

Postby MultiD Support » Fri Jul 04, 2008 9:19 am

Morgan,
If you are comparing the expression ration leaf;root between plants it is a truly paired study, GenEx ver. 4.3.6 has “Normalize with samples”, which is needed for pre-processing of paired data. A beta version of 4.3.6. was recently released. See info under “beta versions” in the Forum.

With only three samples (three varieties) multivariate analysis will not be very exciting. Of course, you can do hierarchical clustering, which will tell you if two of the varieties are more similar than the third. PCA can also support such conclusion. If you obtain such result you can confirm it with mini SOM using a 2x1 network. But clustering is typically more interesting if the number of samples are larger (unless, of course, the goal is to classify genes).

Good luck!
MultiD Support
 
Posts: 170
Joined: Tue Jul 01, 2008 11:49 am


Return to Data Manager

Who is online

Users browsing this forum: No registered users and 1 guest

cron



MultiD Analyses

Home of the GenEx analysis software




Partners






































www.Gene-Quantification.info