The importance of being reproducible. A statistical perspective
24 gennaio 2017
11:00-11:30 Clelia di Serio (CUSSB): Welcome and introduction. The importance of being reproducible: some statistical issues on reproducibility
11:30-12:00 Paola Rancoita (CUSSB) The importance of being reproducible: reproducibility, robustness and validation Coffee Break
12:15-13:00 Claudia Angelini (Relatore Ospite: CNR Napoli) The importance of being reproducible: Challenges and solutions for the analysis of NGS data
Claudia Angelini, IAC, Istituto per la Applicazioni del Calcolo “M. Picone”
NextGeneration Sequencing (NGS) has revolutionized the way of thinking and of performing biomedical research. However, despite the technological advances, the way in which data analysis is performed and described in most of scientific papers does not facilitate the reproducibility of scientific results. Complex analyses are often poorly described in few lines of methods section and they often lacking of the technical information crucial to reproduce the results. Therefore, the problem of the “Reproducible Research”, here-denoted RR, is emerging as a serious issue for all Life Sciences. In this talk, we first introduce the concept of computational RR. Then, we discuss a simple way to develop novel computational tools that implement RR by using R and Bioconductorpackages. In particular, we show how RR can be incorporated within a graphical user-friendly interface and how such tools can automatically generate reproducible analysis. As a concrete working example of our approach, we illustrate RNASeqGUI and IntegrHO, two novel tools that implement RR.