[BioC] saving GRanges objects - resulting file size issue

[Charles Berry]

Valerie Obenchain <vobencha at ...> writes:

> 
> Hi Andrew,
> 
> I can't reproduce this problem. Can you successfully save the files in a 
> different order, i.e., is it just the addition of 'introns' that is 
> giving you trouble?
> 
> I assume these are custom annotation GRanges that did not come from a 
> TxDb in Bioconductor. Can you make the files available for testing?
> 

One thing to watch for is an object in the mcols() or metadata() with
an attr(*, ".Environment") that includes large objects.

formula objects are often culprits in this regard.

Try str() on the GRanges and look for stuff like

attr(*, ".Environment")=<environment: 0x10e28c118>

There were some discussions on R-devel about inexplicably large save()'s
in the last year or so. IIRC, these were the result of saving an object 
containing a formula whose environment (containing some large objects)
was added to the .rda. 


> Valerie
> 
> On 08/16/2013 10:13 AM, Andrew Jaffe wrote:
> > Hey,
> >
> > I'm trying to save several GRanges objects in the same rda file, but for
> > some reason, one of the smaller GRanges objects (~23Mb) makes the saved
> > file go from 25Mb to 9Gb+ and I'm unsure of why exactly, and have never
> > seen this problem before:


[snip]