Coordinate R Versions with Bioconductor

[haydenbspence]

Currently it is necessary to use two versions of R when working with both HADES and OHDSI tooling alongside Bioconductor.

HADES 2024Q1 Wide Release: R Version 4.2.3

Bioconductor Version 3.19: R Version 4.4

• Alignment of R versions would allow for a single R installation that would cover a considerable range of informatics work.

• HADES developers could develop against the Bioconductor devel R version with a reasonable expectation that this would become the next HADES wide release R version (and bio conductor release version).

[schuemie]

We're targeting R 4.4.1 for HADES-wide release 2024Q3. Our current process for selecting the next R version to support is based on balancing site ability to upgrade, and whether the update is required to run HADES. I agree it would be nice to mirror Bioconductor, but not at significant cost to sites to force them to upgrade without a clear need.

Do you know what process Bioconductor uses to determine the R version to support next?

[haydenbspence]

We're targeting R 4.4.1 for HADES-wide release 2024Q3. Our current process for selecting the next R version to support is based on balancing site ability to upgrade, and whether the update is required to run HADES. I agree it would be nice to mirror Bioconductor, but not at significant cost to sites to force them to upgrade without a clear need.

Do you know what process Bioconductor uses to determine the R version to support next?

Bioconductor's devel branch uses different R versions throughout the year

After R's release in April/May until Bioconductor's release in Oct/Nov:

• Bioconductor devel uses the current R release or patched version

From Bioconductor's release in Oct/Nov until the next R release:

• Bioconductor devel uses R-devel

During the R-devel period:

• Builders are updated to the current R-devel version every 4-6 weeks
• This cycle helps Bioconductor prepare for the upcoming R release

[schuemie]

Thanks Hayden!

I think BioConductor's pace of updating is higher than we want it to be for HADES.