[BioC] overlapping genes with biomart

[Sean Davis]

On Wed, Oct 15, 2008 at 7:03 PM, Elizabeth Purdom
<epurdom at stat.berkeley.edu> wrote:
> Hello,
>
> My understanding of the filters in biomaRt "start" and "end" is that they
> correspond to "start>=..." and "end<=..." so if I call
>
> ens <-
> getBM(c("ensembl_gene_id","start_position","end_position","chromosome_name"),filters
> = c("chromosome_name", "start", "end"), values = list(chromosome, x1, x2),
> mart = biomart)
>
> then I would get a return of all genes that completely lie between x1 and
> x2. What I would like is all genes that overlap that region. Ideally I could
> do this by a filter that does "start<=x2" and "end>=x1" but that inverted
> filter doesn't seem to be predefined. Is there a simple way to invert the
> filter? Or am I missing something else obvious?

It is possibly easier to simply pull in all genes and then do the
overlaps in R.  You could also write raw SQL, but that will mean
accessing the mysql server directly and either a temp table (not
possible at ensembl) or multiple queries if you have more than one
region.

Sean