Selecting distance upstream and downstream of TSS for enrichment analysis

[Danielle Hiam]

Hello, 

I have list of 97 DE genes from a RNAseq experiment that I would like to investigate if any transcription factors are enriched based on this list. I would like to use UniBind enrichment function however, I am little unsure about what to include for the BED file. 

Specifically, is there a recommended or consensus of the upstream and downstream distance from a TSS of gene for a TFBS? I have seen 100 upstream and 1000 downstream but also seen many other distances. 

Secondly, is there a function within JASPAR or UNIBIND that can map the TSS of a list of genes? I have tried to use biomart to create the background list  (~15000 genes considered expressed) however am having timeout issues. I apologise if this has been answered before but I am only new to JASPAR and Unibind. 

Any help will be appreciated. 

[Anthony Mathelier]

Dear Danielle,

I'm very sorry for my late reply.

Regarding the extension, I recommend testing different sizes of extension to make sure that the enrichment is consistent over multiple values and make sure it is not biased towards a single set of values.

For the TSSs, I would recommend to use biomaRt and rtracklayer from Bioconductor.