unite histone sites, DNAse, Chip-seq data from different cell types.

Question

[tonja.r]

I am currently running into a problem because I have insufficient amount of the data I need.

I have 81 sets of SNPs and I looked up if those SNPs fall into H3K27ac signal sites.

I am focusing on roadmap data of histone modification (section d (signal tracks), consolidated data, fold-enrichment) of the following cell types: Primary T helper cells PMA-I stimulated (E041-H3K27ac), Primary T helper naive cells from peripheral blood(E039-H3K27ac), Primary T helper 17 cells PMA-I stimulated (E042-H3K27ac).

(I was also wondering where from Fahr et al have RNA-seq data for Tnaive, Th stim and Th17)

My idea is to co-localize my SNPs not only with the histone mark sites but also with DNAse hypersensitivity and TFBS measured with CHiP-seq.

Unfortunately roadmap does not have DNase data of the the cells I mentioned, neither do they have a CHiP-seq.

My questions are: do I really need DNAse and CHiP-seq data of the cells above?

For DNAse I guess I need exactly the data from the cells I look at.

But can I use any CHIP-seq experiment as my goal is only to co-localize my SNPs with TFBS? Or is there any tool that can predict a TFBS from a given DNA sequence?

Answer 1

[inesdesantiago]

The haploreg webtool from the Broad (http://www.broadinstitute.org/mammals/haploreg/haploreg.php) can give you some hints of the TFs that bind to those SNPs. It uses all ENCODE data and the output tells you what are the TFs and cell lines at each SNP. It also use predicted TFBS motifs and their scores to predict if a given SNP can disrupt the binding sequence or not. I find haploreg quite useful for quick checks.

However, I don't think Encode has analysed those cells you mentioned, so, If you are considering a more in-depth and really looking at cell-specific regulatory mechanisms the only way to be more certain would be to use cell-type specific data in that case.