How to get the version history of a bioconductor package

Question

[oganm]

I wan't to see the version history of an R package in bioconductor but it doesn't seem to exist on package main pages. Does anyone know how can I get that information? It would also be nice to be able to get the previous versions of the package

Answer 1

[Martin Morgan]

Might as well ask questions about Bioconductor on the Bioconductor support site.

Bioconductor goes through a formal release cycle every six months. See the 'Previous versions' box on the help overview page to find the version of a package associated with a particular version of R; use that version of R and standard installation procedures (source("http://bioconductor.org/biocLite.R"); biocLite()) to install the corresponding package versions. Alternative, visit the package 'landing page' associated with the version you're interested in to retrieve the source tarball or binary package. Visit the landing page by following the link to the version of interest, e.g., Bioc 2.13) and choosing the package you're interested in, e.g., Rsamtools.

There can be some significant problems with using older versions of R / Bioconductor, e.g., because CRAN installs only the current version of a package, such as RSQLite 1.0.0, a substantial, non-backward compatible revision released a couple of months ago, rather than the historically contemporary version. In addition, the package version available in each release is the last version of the package published for the release, e.g., if version 1.20.0 had a bug that was fixed and appeared as 1.20.1, then only 1.20.1 is still available via biocLite() or as a tar ball or pre-built binary from Bioconductor.

All Bioconductor packages are retained under SVN version control; any version of a package can be obtained via SVN. Again significant caveats apply -- Bioconductor packages are generally compatible within but not between Bioconductor releases, so there isn't much point in hoping that last years package X will work with this year's release; one is better off using biocLite().