Studying Regulatory Elements For Complex Disorders

Question

[Rc]

I am curious to know about the best possible approach to find the regulatory elements upstream of a gene of interest,Which is associated with a complex disorder.

How can we make use of publicly available Chip seq data for this type of studies and what things should be kept in mind,before taking any CHIP data,as There is a diverse range of data,with different cell lines and Some cases,different treatments too,or pooled down with different antibodies.

Hope to get some favorable reply.

Thanks

Answer 1

[Casey Bergman]

Another good way to explore histone and TF chip-seq data is to use data from the Ensembl regulatory build: http://www.ensembl.org/info/docs/funcgen/index.html

Navigate to a gene of interest, such as BRCA2: http://www.ensembl.org/Homo_sapiens/Location/View?g=ENSG00000139618;r=13:32889611-32973805

The click on "Configure this page" on the left hand, which will open a pop-up, in which you should select "Regulatory evidence" from under the regulation option. This will display a matrix of TFs and histone marks arranged by cell type. You can then select multiple cells by clicking-and-dragging across the matrix, or select rows/columns by clicking on the header and "select all"-ing. Once you have chosen your set of regulatory features of interest, click the check mark at the top right to have these tracks be displayed in the main Ensembl "Region in detail" view.

One of the good things about this dataset is that the tracks are all processed using a common pipeline and therefore are (more) directly comparable than sites that combine data based on variable mappers/peak callers.

[Khader Shameer]

Nice approach Casey, thanks a lot for sharing this.

[Larry_Parnell]

I agree - nice approach as well.

Answer 2

[Khader Shameer]

IMHO, we don't have a standard approach for regulatory genome analysis. AFAIK, at the moment it is custom-designed for the disorder, gene or tissue of interest.

I would recommend this excellent review article for a background reading.

Also if you are specifically interested in Encode data, this how-to style article written by the authors with free ENCODE / UCSC Genome Browser tutorials at OpenHelix will be very useful.

[Rc]

Thanks a lot khader.Sorry for delay in response.

Answer 3

[Larry_Parnell]

I think this is a direction that the Pritchard lab at U Chicago is taking. What you are really talking about is the intersection of eQTL data, disease association (genetic, medical or other), and DNA motif analysis. Yes, this is cell type specific, but data for a similar or any cell will be of value because it'll show something about the gene's behavior. I would take a look at http://eqtl.uchicago.edu/Home.html and things like CENTIPEDE (inferred transcription factor binding sites) and the eQTL data, whose shared sites on the genome really begin to describe where you want to take your research.

And, a +1 for Khader as the ENCODE et al projects will give a lot of useful ChIP data about regulatory regions.

[Rc]

I do agree with you Larry,This is what I intend to do with my work,But It seems to be a bit complex.well I would really appreciate your efforts and sharing the knowledge.

[Larry_Parnell]

With "a gene of interest" you may really have a couple: that gene plus a couple nearby neighbors. I'd begin by using the old SymAtlas data to learn in which tissues/cell types these genes are expressed - any relevant to the disease? Then, look for motifs for TFs that are known to be expressed in those tissues or use ChIP data for those tissues and not others. In other words, give this focus by looking at data only from the relevant cell type or tissue.

[Rc]

That is of-course a nice approach,I think the best possible to make use of the available datasets.

Answer 4

[Asaf]

I would start with regulation tracks in the UCSC genome browser (make sure you use hg18) or ensemble, you can find data from a lot of ChIP-seq experiments there.

[Rc]

Thanks Asaf,But the chip seq data is for different cell lines and that doesn't mean that the regulation is similar in all.

This is my confusion

[Asaf]

Maybe this paper will help you (or increase you frustration): http://www.ncbi.nlm.nih.gov/pubmed/21473766 it claims that DNA accessibility governs the TF binding so you should look for accessible regions first. This one is also relevant: http://www.ncbi.nlm.nih.gov/pubmed/21304941

[Rc]

Thanks Asaf,Its getting more complex to think that way.Well would try to gain the insight of it.

Thanks a lot