Improving the computational reproducibility of immunological research

Abstract

Abstract Many excellent and free training resources exist to improve the computational reproducibility of biomedical research. However, most of these materials target researchers at the stage of data analysis, and they typically provide much less guidance on principles and techniques to improve the reproducibility of the earlier steps of experimental data recording and experimental data pre-processing that are critical for ensuring the reproducibility of immunological research. Many common existing practices---including the use of spreadsheets with embedded formulas that concurrently record and analyze experimental data, problematic management of project files, and reliance on proprietary, vendor-supplied point-and-click software for data pre-processing---interfere with the transparency, reproducibility, and efficiency of laboratory-based immunological research projects. Under an NIH grant, we are creating new training modules as a resource for laboratory-based biomedical researchers to learn how simple computational principles can improve reproducibility at the stages of data recording and pre-processing in immunological research. In this poster, we will highlight several simple approaches that we have identified through this project to improve the reproducibility of immunological research, including strategies for re-designing data collection spreadsheet templates to more easily integrate into a reproducible data analysis workflow and the use of reproducible reporting tools within a laboratory setting.