Small genomic region proximity enrichment

Question

[Nicolas Rosewick]

Hi,

I've a bunch of very small genomic regions (< 1kb) that seems to be non-randomly distributed into the genome. Anyone has an idea to test whether these regions are close together ? or some of them are clustered in specific genomic regions?

For the first question, I was thinking to use a kolmogorov-smirnov test to test if the small regions coordinates are uniformly distributed across the genome.. any advice?

Edit: I'm working in human (~70 regions of less than 1 kb)

Thanks

[Ram]

Hi,

The following details might help us help you better:

1. The organism you are working on
2. Details on the genomic region data you already have

Thank you!

[Nicolas Rosewick]

I edited the main post

[Ram]

Thank you. Could you also add details on the data - you have given us the biological premise, but not the structure of the data. What do you have - only the sequence or any mapping details (such as chromosome to which these regions map or any genes these regions are linked to). That would help too!

[Nicolas Rosewick]

I've the sequences, and the associated alignment (blat)

[Devon Ryan]

How did you find the regions originally? The tricky part to deal with is that if this was generated by sequencing then you wouldn't expect regions to be randomly located due to various biases (GC content, mappability, etc.).

[Nicolas Rosewick]

yes my data originate from RNA-Seq data. It's a good point to take that into account

Answer 1

[Zev.Kronenberg]

There is a really good R package for testing these sorts of questions: http://genometricorr.sourceforge.net