High throughput microscopy

Modern whole-slide microscopy scanners are able to digitize whole-brain tissue sections from human brains at high throughput. Our institute produces multiple Terabyte of image data from different types of microscopes each day. To deal with this large continuous stream of data, we develop distributed data management workflows which transfer images safely from our lab to storage resources at the Jülich Supercomputing Centre (JSC) perform automatic quality control and image analysis on JSC’s compute clusters, and allow remote inspection of all steps in the pipeline for scientists in the institute. Important aspects of this work are provenance tracking, metadata curation, search indexing and remote visualization. We aim to set up a publicly accessible data repository, and interact with the biomedical imaging community on the development of international standards. We work closely with the developers of datalad and UNICORE, and are involved in the development of the openMINDS metadata standard for neuroscience data within the Human Brain Project (HBP).