Hi,
Does anybody knows that genes expression are linear or non-linear ? I want to find out co-expressed genes in microarry study data and people usually use correlation to find out co-expressed genes. I wounder, whether genes expression with identical TF are linear or non-linear ? (cuz, correlation just detect linear relationship between two variables)
There are some studies on which measure to use to correlate genes or patient samples, for instance:
Pablo A. Jaskowiak, Ricardo J. G. B. Campello, Ivan G. Costa Filho, "Proximity Measures for Clustering Gene Expression Microarray Data: A Validation Methodology and a Comparative Analysis," IEEE/ACM Transactions on Computational Biology and Bioinformatics, vol. 99, no. PrePrints, p. 1, , 2013
Note that Pearson only detects linear correlations, but if you use, for instance, Spearman, you can catch some non linear correlations as well.
Hope it helps, Cheers
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
Identical what transcription factors? The same binding sites? The same signal from a ChIP-seq experiment? Something else entirely? Biological effects are rarely linear.
good, thanks, so what is good criteria instead of correlation analysis for finding co-expressed genes in microarray study ?
I remember there being a neat method introduced in PNAS a few years back, but I expect I'll have to flip through the stack of papers in my office to find it (I'm on vacation until next week, so it could be a while). BTW, keep in mind that just because things are often not really linear, that doesn't mean that treating them as linear won't still work (sometimes it will, sometimes it won't).
thanks dpryan. one more thing. I have buch of micRNA genes, but for 85% of them I don't have TF. is there any tools or webserver which predict the TF of micRNA ?
I see that you got a reply in your other thread. Otherwise, there are a number of them out there if you search around.