Questions and Answers for new Bioconductor package authors

Marc Carlson, Martin Morgan, Sonali Arora
December 10th 2014

Before we begin:

  • Welcome to the 1st Bioconductor Q&A!
  • Please post any questions into the YouTube chat window
  • After after a short talk, I will be able to answer them

Major goals for the Bioconductor project:

  • Analysis & comprehension of high-throughput genomic data
  • Robust software
    • Documented
    • Tested
    • Freely Available
  • Reproducible research
  • Code reuse and the importance of working together

What is package review for?

  • To improve your package
  • To ensure that your package helps meet the goals of the project
  • To help you to know more about other existing packages/objects

Resources for new package authors:

What do reviewers look for?

source("http://bioconductor.org/biocLite.R")
biocLite("BiocCheck")
library(BiocCheck)
  • Code reuse
  • Good documentation - Everyone must understand vignette.
  • Why do people care about how my code looks? (code readability)

Should I publish my academic paper or my package 1st?

  • Our svn codebase is public
  • Many journals like if your software is already online
  • Ultimately: you have to check with the specific journals (It's their policies that matter)

What to do when it's time to submit your package:

  • Don't panic - the process is meant to help you.
  • Build and check your package (using Bioc ' devel'!)
R CMD build MyPkg
R CMD check MyPkg_0.99.0.tar.gz
R CMD BiocCheck MyPkg_0.99.0.tar.gz

Acknowledgments

  • Bioconductor is supported in part by these grants:

    NHGRI / NIH U41HG004059

    NSF 1247813