Associating Gc Content With Genomic Features

Question

[Gvj]

Hey all,

When I have identified certain events in genome and would like to check whether it is associated with GC content, histone mark etc, What kind of statistical test I should do ?

Eg: I have calculated the GC content (y axis) from the event till 1kb. Now knowing the average GC content as 0.36 in the genome, How can i test whether my events are associated with GC less regions ( may be the first 100 bps )

Answer 1

[Sean Davis]

A simple answer is to make a histogram of the GC content for each of your regions. Then, look at where your genomic GC content falls on the histogram. If it falls near one of the tails, you have evidence that your regions are different from the genome.

The "real" answer question depends on what hypothesis you want to test. A typical approach would be to choose an "equivalent sample" of regions from your genome of interest and compare the GC content of these "equivalent regions" to that from your sample using something like a t-test. Choosing a set of "equivalent regions" can be a bit challenging, but you could start with random regions and get more complicated if necessary.