Multiple inheritance might work. May be a good time to think about an
abstract "sequence source", i.e., anything that supports getSeq(). Right
now, we use ANY for that, which is not ideal. Same for "transcript source".
We've already implemented the TxDb API directly on top of biomaRt.
On Wed, Ju
Sounds nice but we have to be careful. We may need a parameter:
Homo.sapiens("hg19") ... because
broad adoption of new builds may take considerable time.
On Wed, Jun 18, 2014 at 6:13 PM, Hervé Pagès wrote:
> On 06/18/2014 05:26 AM, Michael Lawrence wrote:
>
>> Would it make sense for an Orga
On 06/18/2014 05:26 AM, Michael Lawrence wrote:
Would it make sense for an OrganismDb object to have an associated
BSgenome? This association is implicit anyway. Or do we need an even
broader umbrella object? This would make some things easier. For example,
predictCoding(variants, Homo.sapiens).
Would it make sense for an OrganismDb object to have an associated
BSgenome? This association is implicit anyway. Or do we need an even
broader umbrella object? This would make some things easier. For example,
predictCoding(variants, Homo.sapiens). Just nice to have everything in one
bundle. Also,