Creating a Biomedical Dataset: Conceiving a Purpose Statement, Developing an Experimental Protocol, and Addressing Data Management
摘要
Starting the journey of creating a biomedical dataset can be a daunting task, particularly for the inexperienced researcher. This chapter provides actionable guidelines to achieve a representative, reproducible, and shareable experimental study. Without extensively addressing how to design a study plan, this chapter summarizes the core requirements that should be assured as a proper foundation for the experimental design. The current regulations for legal and ethical compliance are covered, including the necessary documents that must be submitted for approval. This chapter also provides a comprehensive description of the principles behind an experimental protocol, targeting a clear narrative, minimizing potential issues, and expediting problem resolution. Finally, dataset management is addressed, providing the reader with a set of relevant features and factors to consider during the choice of file formats and dataset structuring. Together with the guidelines, this chapter provides an application to a real-life case study, namely, the H2A2 Innolabs experiment—a study performed collaboratively by the company VicarVision and the research group Pattern and Image Analysis-Lx.