Michelle Itano, Ph.D.
Enhancing Reproducibility Through Quality Control, Date Management, and Imaging Analysis with Collaborations Across Disciplines
Currently, researchers are forced to manually track and annotate all imaging experiment steps, from sample preparation, to image data acquisition, storage, processing, and analysis, which is very onerous, slow, and wasteful. Lack of automation and poor documentation often makes it impossible to reproduce, compare, and combine large sets of data from different laboratories, biological systems, or different experimental setups, making it difficult for scientists to collaborate with each other. Our Light Microscopy Core, UNC Neuroscience Microscopy Core, has begun to address these issues through productive collaborations including other facility managers, computational scientists, individual researchers, and IT experts. One overarching collaboration is focused on how to improve the data infrastructure between researchers and Core Facilities to minimize data duplication and slow transfer rates and maximize computational resource usage for storage and analysis. This has led to multiple additional projects: to support more consistent and automatically generated image metadata and calibration control records; to implement data storage organized by research project including data generated from multiple methodologies (e.g. microscopy, MRI, mouse behavior); and to provide analysis projects that span multiple integrated methodologies (e.g. sub-second 3D mouse behavior image analysis). While still in initial stages of development, these preliminary projects suggest that Core Facility collaborations across disciplines can be harnessed to address some of the most critical issues in our field.