[ensembl-dev] Can you link TF-binding sites to the genes they are predicted to regulate in the funcgen database?

Ian Dunham dunham at ebi.ac.uk
Fri Feb 17 15:04:54 GMT 2012


We're kind of wondering this too :-)

Seriously this is significant aim for a lot of people in the field, and 
we'd like to be able to provide this through Ensembl.  There are 
different ways to do it, some of which are achievable now but which need 
careful interpretation, and others that will require good validated 
data. A brief synopsis of possibilities includes

1. A simple method that is used in the field is to calculate the nearest 
gene, or more precisely the nearest 5' end of a gene to the TF binding 
site.  This assumes that the TF is likely to regulate a promoter closest 
to it, which is frequently but not always the case.  One can implement 
an approach to do this, but we haven't explicitly made this link on the 
grounds that it could be misleading since the biology may not always 
behave that way.

2. Interaction mapping such as the various flavours of 3C, 4C, 5C and 
ChIA-PET should give sets of interactions between regulatory sites that 
we could use to imply regulation.  Interactions between promoters and 
distance sites in a cell found in this way would implicate regulation of 
that promoter in that site.  There are complications because it is 
possible for cell specific differences.  We're looking for good 
validated data that we can use in this way, we are aware of at least 
some data that could be used but there are caveats. Other methods that 
map looping interactions would also work here.

3. Association analysis between signal intensity and, say, expression 
level by RNA-seq can also be used to infer a relationship between sites 
and genes.  There is data from several sources now that gives some of this.

At present our feeling is that neither of the second two is mature 
enough yet to implement within ensembl but we are monitoring this area, 
because it is something we want for several reasons.

We'd be interested in any other suggestions that are out there. I may 
have neglected some approaches in this brief reply.


Ian Dunham

On 17/02/2012 14:50, Mark Aquino wrote:
> Hi all,
> I was wondering if it was possible to map the positions where binding factors are predicted to bind (in the funcgen database) to the genes they are predicted to regulate.  If anyone knows how to do this I would be very appreciative.
> Best,
> Mark
> _______________________________________________
> Dev mailing list    Dev at ensembl.org
> List admin (including subscribe/unsubscribe): http://lists.ensembl.org/mailman/listinfo/dev
> Ensembl Blog: http://www.ensembl.info/

More information about the Dev mailing list