Reconstructions and analyses from the IARPA MICrONS consortium mouse cortex Layer 2/3 serial EM volume.
Attribution: https://www.microns-explorer.org/terms-and-conditions
Citation: https://www.microns-explorer.org/citation-policy
For an intriguing summary of this revolution, read Mitch Glickstein's essay on Golgi and Cajal in Current Biology.
Image credit: I created this serial LM reconstruction of a Golgi-stained mouse pyramidal neuron using a Zeiss Photomicroscope II at 40x oil and Smart Objects in Photoshop.
Once again neuroscience is undergoing another revolution with the serial electron microscope brain reconstructions
Petabytes of data have been generated for small volumes (1 cubic mm or less) of human, mouse, songbird, and fly brain volumes. This gargantuan task has been spearheaded by investigators from a variety of entities, including Allen Institute, Baylor, Columbia, Google Research, Harvard, IARPA, Janelia/HHMI, Johns Hopkins, Max Planck Institute, Princeton, Rice, and the University of Cambridge, among others.
Image credit, I created this reconstruction using Neuroglancer of pyramidal neuron with cellid 648518346349538440 in the Layer 2/3 volume.
This github repo focuses on visualization and analyses of the Layer 2/3 EM volume data generated in the IARPA Microns consortium (Allen Institute, Baylor, Princeton). The volume can be viewed in Neuroglancer (developed at Google Research). Allen Institute and the Seung lab at Princeton have posted datasets, resources, analysis tools, and more on their respective github repos, and also at the Allen Brain Map website.
Use lookup_cellid_in_layer23_volume_neuroglancer.ipynb to identify the cell subtype for a given cellid in the Layer 2/3 volume. This is a human-curated list for cells that contain all (or most) of the cell soma within the volume. Includes a Neuroglancer link generator. See also a version that adds the Neuroglancer nucleus id (for cells with a nucleus in the volume): lookup_cellid_in_layer23_volume_neuroglancer_with_nucleus_id.ipynb.
Use list_of_cell_subtypes_in_layer23_volume.ipynb to view the cellids categorized by cell subtype, from a curated list of 619 cellids.
Use lookup_mitochondria_ids.ipynb to look up mitochondria ids by cellid. Useful for creating a list of mito ids for visualization in neuroglancer and for analyzing voxel data.
Use neuroglancer_link_generator_mitochondria_visualization_version.ipynb to generate a Neuroglancer link for a single cellid and single mitochondrion from the pni_mito_cellswskel_v185_fullstats.csv datatable.
Use neuroglancer_link_generator_all_mitochondria.ipynb to generate a Neuroglancer link for all the mitochondria in a single cellid of interest pulling the mito ids from the 211019_mitochondria_info.csv datatable.
Use vtk_mitochondria_visualizer.ipynb to generate a 3D interactive view of all mitochondria in a cell of interest. Note this method has significant file download and space requirements compared to using the Neuroglancer viewer. Use vtk_mitochondria_visualizer_with_synapses.ipynb to add the pre- and post-synaptic sites to the visualization as well.
Use lookup_synapse_ids.ipynb to look up synapse ids by pre- and post-synaptic cell ids. Also contains voxel data.
Use synapse_visualizer.ipynb to create a 2D and 3D visualization of all pre- and post-synaptic sites on a cell of interest.
synapse_visualizer_plotly.ipynb is an updated synapse visualizer with interactive 2d and 3d plotly scatterplots added.
Use vtk_nuclei_visualizer_with_vasculature.ipynb to create an interactive 3D visualize the cell nuclei in the volume using vtk and OpenGL. Includes a visualization option to include all vasculature in the volume.
Use layer23_cell_proximity_calculator.ipynb to enter a cell id of interest and make of list of nearest cells, using a simple centroid-to-centroid Euclidean distance calculation.
axon_carrying_dendrite folder: a partial neuron (soma is not in the volume) with a possible axon-carrying dendrite.
autapses folder: there are three autapses in the Layer 2/3 volume and many (24) cases of segmenation errors that are not true autapses.
mitochondria folder: analysis of interesting mitochondrial features in the layer 2/3 volume, including the largest contiguous mitochondrion in an astrocyte, and the largest number of mitochondria by count in an inhibitory basket neuron.
decimated meshes folder: notebooks and vtk visualizations using pyvista mesh decimation to reduce the size of the cell body and mitochondria mesh files.Modified from Tyler Sloan's() mesh decimation pipeline.
astrocyte mitochondria folder: analytical and visualization notebooks and high resolution 3D renderings of astrocyte mitochondria.
astrocyte mitochondria inclusions folder: visualization notebooks for generating electron micrograph (EM) images and high resolution 3D renderings of mitochondria inclusions in astrocytes of the Layer 2/3 volume.
mitochondria analytics folder: more in-depth analysis of the structural, spatial, and positional characteristics of mitochondria in the Layer 2/3 volume.
most_synapses folder: reconstructions and synaptic analyses of an excitatory pyramidal neuron with the most synapses in the volume, along with an inhibitory basket neuron making the most synapses onto other processes within the volume.
Pairs of neurons in the Layer 2/3 volume that synapse onto one another
reciprocal_pairs folder: Martinotti-bipolar reciprocal pair (cell ids 648518346349538179 and 648518346349515986, respectively).
oligodendrocyte_648518346349508279 folder: a beautiful example of an oligocyte in the layer 2/3 volume. Images of myelination as well as an example of mis-segmentation where a synaptic bouton from a nearby axon was incorrectly segmented to the oligodendrocyte cell.
vtk_nuclei_visualizer_with_vasculature.ipynb: notebook to create a 3D visualization of cell nuclei in the Layer 2/3 volume. Uses meshparty, vtk, and OpenGL. Here, nuclei from excitatory neurons is shown in light blue; nuceli from inhibitory neurons in light red, and nuclei from glial cells are shown in olive green. Vasculature in the Layer 2/3 volume is shown in light grey.
Drop me an email if you have any questions or would like to collaborate. Shawn