An R Package That Collects and Archives Files and Other Details to Support Reproducible Computing

• Activity Number: 84
• Type: Contributed
• Date/Time: Sunday, August 9, 2015 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistical Computing
• Author(s): Stan Pounds* and Zhifa Liu
• Companies: St. Jude Children's Research Hospital and St. Jude Children's Research Hospital
• Keywords: reproducible computing ; R package ; scientific method ; trace function ; data analysis archive

Abstract

Reproducible computing is an essential element of the scientific method in the digital era. Published data analysis results that are not computationally reproducible are not fully transparent for peer review and independent replication by other investigators. The R package rctrack uses the trace function to embed detail collection procedures into the definition of functions that access files, generate random numbers, and other computational processes that must be documented to reproduce a data analysis at a later time. By directly embedding these detail collection procedures into the definitions of relevant R functions, rctrack eliminates the burden for users to make special modifications of existing programs to make analyses reproducible. Furthermore, upon completion of the analysis, these files and details are automatically copied into a special time-stamped read only archive directory. The package is illustrated with some real example data analyses from statistical cancer genomics applications.