Selecting Urn/Universe For Hypergeometric Test

Question

[Misty]

Hello,

I am trying to analyze dinucleotide representation in TF consensus sequences in human promoter regions. However, I can't decide whether to use only the promoter sequences or the entire human genome as the "universe/urn". Any suggestions?

I am using R for my analysis and would also like to know if there are any packages that can retrieve the promoter sequences for a list of TFs. I am using BSgenome (BSgenome.Hsapiens.UCSC.hg19) and GenomicFeatures but they require I specify the TFs one at a time.

Any comments/suggestions are greatly appreciated!

Thanks!

[Sean Davis]

With regard to your second question about BSgenome and GenomicFeatures, I'd suggest you post the code you are using. As to the first question about "background", that is a decision only you can make as the answer depends on the hypothesis you are testing. Perhaps you can clarify what your hypothesis, specifically, is.