1. About
  2. Features
  3. Explore

I have a new academics R-package for bioinformatics and I would like to submit my package online.

  1. Where I should submit? If I submit to Bioconductor, they have a 6 months release cycle. What about CRAN? Will they release my package once it's reviewed?

  2. I don't think the users care where they get the package from; CRAN or Bioconductor. What's the advantage of submitting to Bioconductor? Why not just to CRAN? Why all the bioinformatics packages get submitted to Bioconductor? I don't see they actively promote the packages. Why not to CRAN?

  3. Can I submit to CRAN, then later change to Bioconductor?

  4. I have a very detailed user guide in DOC/PDF format. However, the Bioconductor requires a running vignette. I think they mean a markup document that all the examples can be run automatically by the build process. What should I do? Should I convert the 100 pages guide into the markup format? Should I just submit to CRAN because I don't think (but I'm not 100% sure) they require a vignette.

  5. Does a static DOC/PDF document sufficient as a vignette?

1 Answer 1

There's a lot of questions here and I'll try to address them. Full disclosure: I'm experienced with CRAN but not with developing on Bioconductor.

  1. Where you should submit depends on a variety of factors. Since it's a bioinformatics package, it would fit well in Bioconductor, but also in CRAN. A CRAN maintainer will review the contents of your package for compatibility with CRAN's standards and let you know about any issues. If there are no issues, the package can be published on CRAN within a few hours of submitting. Though sometimes it takes longer. I've never had to wait more than 24 hours.

  2. I'm not in bioinformatics, but my best guess is that Bioconductor exists to help make finding relevant (i.e. bioinformatics) packages easier. According to the website Bioconductor provides tools for the analysis and comprehension of high-throughput genomic data. Thus it may be advantageous to put your package there because it is more likely to be found. CRAN currently has 8907 packages available while Bioconductor only has 1213 software packages. While this is still likely too many to sift through, it may be easier for users to find your package than amongst all the non-bioinformatics packages also on CRAN.

  3. According to Bioconductor's package guidelines page: Authors are strongly discouraged from placing their package into both CRAN and Bioconductor. This avoids burdening the author with extra work and confusing the user. If you plan on submitting to Bioconductor, I'd suggest going straight there and avoiding CRAN altogether. It's frustrating to the end-user to write lots of code using one package to find that the package is changed and your code is no longer compatible and needs to be manually fixed everywhere you used it.

  4. CRAN does not require vignettes. Changing a Word document to a Markdown language may be faster than you expect and worth the effort if your functions require explanation for most users.

  5. Static documents won't fly on CRAN as a vignette. Here's some good info from Hadley Wickham: http://r-pkgs.had.co.nz/vignettes.html#vignette-cran

My suggestion is to submit to Bioconductor. Your audience is more likely to find it there. If you're concerned about the wait to publish, you can make a public repository on GitHub where your peers can find your package.