科学研究

Seminars

Industrial Level Full Neuron Morphology Screening of Whole Brains

2019-07-09

Hanchuan Peng,PhD

Director

SEU-ALLEN Joint Center; Affiliate/Adjunct Professor,

University of Washington (USA), University of Georgia (USA), Southeast University (China);

Fellow,

American Institute for Medical and Biological Engineering


Seminar   Type

B-type



Preferred   Location

First Floor Lecture Hall, Jianzan Building (Phase   I)

Chinese Institute for Brain Research,   Beijing


Time

10:30-11:30 am Wednesday,July 10, 2019


Host

Dr. Minmin Luo



Topic

Industrial   Level Full Neuron Morphology Screening of Whole Brains


Abstract

Despite substantial advancement in the automatic   tracing of brain cells' 3D morphology in recent years, it is challenging to   apply existing algorithms to very large image datasets containing billions or   more voxels, especially for applications such as morphometry of single   neurons at the whole brain scale. We have developed a new platform combining   several newly developed technologies including Vaa3D (Nature Biotechnology   2010; Nature Protocols, 2014), TeraFly (Nature Methods, 2016), UltraTracer   (Nature Methods, 2017), and TeraVR, to attempt this challenge. Particularly,   we have used TeraFly to invoke Vaa3D to quickly visualize the whole mouse brain image   volume and manage the thousands of billions of voxels in each of the brain   volume. We then used UltraTracer to wrap several   efficient base tracers to trace such massive data volumes. Finally, we   combined virtual reality and machine learning into a tool called TeraVR for   efficient proofreading and editing of such reconstructed neuron morphology.   We are further improving the integration of these tools for more scalable and   accurate single neuron morphometry.


Speaker Biography


Hanchuan Peng develops revolutionary technologies to   generate, manage, visualize, analyze, and understand massive-scale structure   and function data related to brains. Peng also led the Big Image Mining team   at Janelia, HHMI (2006-2012). Peng is a highly cited inventor of a number of   new algorithms and software/hardware systems, including Vaa3D - a widely   adopted high-performance platform for very large multi-dimensional images   (Nature Biotechnology 2010; Nature Protocols, 2014; Nature Communications,   2014, 2019; Nature Methods, 2016, 2017), BrainAligner (Nature Methods, 2011),   SmartScope (Nature Communications, 2014; Scientific Reports, 2017), mRMR   (IEEE-TPAMI, 2005; a top-10 most cited paper in TPAMI since 2005), etc. He   built/co-worked the first digital maps for several widely used model systems   at single cell/neurite resolution (C. elegans - Nature Methods, 2009; Cell,   2009; fruitfly - Nature Methods, 2011; mouse – Nature 2014; Nature   Neuroscience 2019), and led the “BigNeuron” initiative (http://bigneuron.org;   Neuron, 2015). Peng was inducted into AIMBE (2019), a co-recipient of USA   National Academy of Sciences’ Cozzarelli Prize (2013), DIADEM award (2010),   SfN’2018 Hot Topic (2018), etc. His work has been featured in Nature,   Science, NPR, NBC, etc. Peng founded Bioimage Informatics conferences (http://bioimageinformatics.org), helped iconize   Bioimage Informatics as a new field in major bioinformatics journals   including Bioinformatics, BMC Bioinformatics, Nature Methods, Nature   Biotechnology, etc. He currently serves as the co-Editor-in-Chief of the   journal Brain Informatics (2016-) and associate editors of a number of   journals; he just retired from the Section Chief Editor of BMC Bioinformatics   (2011-2018) overseeing the section of imaging, bioimage analysis, and data   visualization