Identifying Cis- And Trans- Factors

Question

[Raygozak]

Basically i'm doing an analysis on a paper that finds patterns of cis- and trans- regulatory factors, so far i've made the differential expression analysis which consist of analyzing the gene expression on a hybrid of two parental strains. Does anyone could guide me how to identify cis-trans factors for the differential expression profile?

thanks

[Casey Bergman]

Do you have genotype information for the P's and F1? Do you have parent-of-origin allele-specific expression information for the F1?

[ff.cc.cc]

If you have genotype data have a look at http://eqtl.uchicago.edu/Home.html. They host datasets and good references about eQTL and gene regulation

Otherwise restrict your investigation to a more limited target (e.g. a specific pathway) to limit issues due to false positives and multiple testing. You need a strong biological or experimental prior to support the hypothesis that sets of differentially expressed genes should also be co-regulated

Answer 1

[David Quigley]

I'm going to assume that you're looking to identify genes that are influenced by alleles that differ between the parents. If you have gene expression for the parental strains and a F1 hybrid you can identify candidates for genes that may be under genetic control. For example, for strains A and B, genes where there is a good for for a linear model between number of B alleles and gene expression are plausible candidates.

However, you can't determine whether these effects are due to locally-acting factors (e.g. promoter variations in the gene itself) or distant trans-acting factors (e.g. variations in constitutive activity of a transcription factor that affects the gene under consideration). There also might not be a direct genetic effect at all. To get more information there you'll need to perform an eQTL linkage study where you model the effects of allelic variations on a backcross, intercross, RI, or some other controlled population.

Beware of the possible effects of strain-specific bias in the microarray platform if this is a microarray-based study. In short, polymorphisms that affect reported expression levels may not be distributed equally between your strains, and this bias may cause an artificial inflation in the number of apparent cis-acting genetic effects.