Petabyte-Scale Multi-Morphometry of Single Neurons for Whole Brains

Abstract Recent advances in brain imaging allow producing large amounts of 3-D volumetric data from which morphometry data is reconstructed and measured. Fine detailed structural morphometry of individual neurons, including somata, dendrites, axons, and synaptic connectivity based on digitally reconstructed neurons, is essential for cataloging neuron types and their connectivity. To produce quality morphometry at large scale, it is highly desirable but extremely challenging to efficiently handle petabyte-scale high-resolution whole brain imaging database. Here, we developed MorphoHub to address this challenge by optimizing both the data and workflow management. In particular, this work presents a multi-level method to produce high quality somatic, dendritic, axonal, and potential synaptic morphometry. Our method also boosts data sharing and remote collaborative validation. We applied MorphoHub to a petabyte application dataset involving 62 whole mouse brains, and identified 50,233 somata of individual neurons, profiled the dendrites of 11,322 neurons, reconstructed the full 3-D morphology of 1,050 neurons including their dendrites and full axons, and detected 1.9 million putative synaptic sites derived from axonal boutons. Analysis and simulation of these data indicate the promise of this approach for modern large-scale morphology applications.