[Bioc-devel] looking for advice on best R packages implementation strategy

[Julien Wollbrett]

Hello,

I am looking for advice on the best strategy to follow when developing 
several functionalities in R packages for the same resource.

I am developper/maintainer of 2 Bioconductor R packages (BgeeDB and 
BgeeCall). These packages are developed in the context of Bgee, a 
database for retrieval and comparison of gene expression patterns across 
multiple animal species.

* BgeeDB R package allows to:
- download/filter processed Bgee gene expression data
- run a GO-like enrichment of anatomical terms, mapped to genes by 
expression patterns. It is similar to topGo but using the Uberon 
anatomical entity ontology rather than the GO one.
* BgeeCall allows to use Bgee to automatically generate present/absent 
gene expression calls on your own RNASeq data.

We are thinking of implementing new functionalities in R and don't know 
which approach to use :
- reshaping our R packages to have one functionality per package (e.g : 
download data, topAnat enrichment, generation of calls using user data). 
Then, create one new package for each new functionality. This approach 
could be combined to the creation of an helper package loading all Bgee 
packages.
- reshaping our R packages to create one big R package containing all 
functionalities. New functionalities will then be implemented in this R 
package.
- keep our two R packages. Add new functionalities to already existing 
packages or create new R packages

Do you have any advice based on your experience about the best strategy 
to follow?
Thank you for your feedbacks.

Best Regards,

Julien Wollbrett