I'm trying to obtain a list of gene coordinates for mouse/human where all isoforms have been collapsed into one single representative "gene".
I'm primarily interested in using this for generating ChIP-seq binding profiles across genes. If I use annotations with multiple isoforms, this can lead to problems with data duplication (e.g. a single region of enrichment is represented multiple times due to multiple isoforms).
This is just for rough work when I would like to do things quickly without having to sort out and isolate the various different isoforms.
I have looked in the UCSC table browser but I couldn't locate any annotation that looked like this. I have previously tried merging the isoforms myself but I would be more comfortable working with a reference set rather than my own custom set.
Do you want the individual exons collapsed or just the end-to-end gene coordinates? If is is the latter you can simply use the coordinates from all the lines the have feature type "gene" in the gencode annotations, that is available both for mouse and human in gtf format at http://www.gencodegenes.org/
That's perfect. Thank you!