Topic: Data Science
Podium Presentation in Room 3 on Wednesday at 9:00 (Chair: Randall Julian)
Authors: Olga Vitek
Quantitative mass spectrometry-based proteomics aims to distinguish systematic variation in protein abundance (due, e.g., to a treatment or a disease) from nuisance biological and technological variation. Statistical mindset is key for doing so in both repeatable and reproducible manner. Frequently, statistical tasks are viewed as limited to detecting differentially abundant proteins. In reality, statistical components of reproducibility are substantially broader. They include all aspects of data processing (Which features should we use to quantify a protein? How should we combine the features into a protein-level conclusion?). They also include aspects of experimental design, from both biological perspective (Which proteins and samples, and how many, do we need to quantify?) and technological perspective (Are the assays appropriate for the task? Do the experimental steps run properly?). Answering these questions requires the availability of statistical methods, and but also of publicly available data that help understand the advantages and the limitations of the methodological choices. This talk will highlight the contributions of our lab to these components of reproducible research.
|Planning to mention or discuss specific products or technology of the company(ies) listed above:||