Install a specific release of bioconductor package

2019-06-27 10:29发布

问题:

I had a bioconductor package version 1.2.14, and I've just updated it to version 1.4.2 and unfortunately I get different results.

I want to make sure this is because of the package update. Is there a way to revert to the old version, or alternatively to remove the current version and specify installation of the old one?

Alternatively, I can see the branches on GitHub and I wonder whether they correspond to the different version from Bioconductor. So, can I install from a GitHub branch release?

回答1:

Step1 In Object Browser (of Revo R Ent), right-click the bioconductor package under "Packages Loaded", and then click Unload Package.

Step2 In Object Browser (of Revo R Ent), right-click the bioconductor package under "Packages Installed", and then click Unload Package.

Step3 Close the R's GUI (in my example:Revo R Ent).

Step4 (For Revo R Ent users) Delete the .xml file:

C:\Users\User\Documents\Revolution\32_7.3\PackageXMLs\bioconductor.xml

The file bioconductor.xml (related with v1.4.2) has the duty of the functions of bioconductor 1.4.2 to be processed in R's GUI.

Step5 Get your library location via:

.libPaths()    
[1] "C:/Revolution/R-Enterprise-7.3/R-3.1.1/library"

Step6 Delete the folder of bioconductor package from Windows Explorer in the library location. (Now, you got rid of bioconductor 1.4.2)

Step7 Open Revo R Enterprise. Install bioconductor package version 1.2.14: Install - Install Packages from Local Zip Files...- Select .zip of bioconductor 1.2.14 - OK.

Now, bioconductor folder (related with 1.2.14) is listed in your library location.

Step8 library(bioconductor) # Now, bioconductor v. 1.2.14 works in your R's GUI.

(For Revo R Ent users) The new file bioconductor.xml (related with v1.2.14) are created in C:\Users\User\Documents\Revolution\32_7.3\PackageXMLs

Even if you close Revo R Ent, this .xml file sustains in the PackageXMLs folder to act as bufferly.

Though I gave the solution for "Revolution R Enterprise" GUI for R, the solution mechanism is the same for other R's GUIs.