neuPrint#

This tutorial shows how to fetch neurons from a neuPrint server.

NeuPrint is a service for presenting and analyzing connectomics data. It is used to host, for example, the Janelia EM reconstructions from a Drosophila hemibrain at https://neuprint.janelia.org/.

neuprint-python is a Python library that lets you query data directly from a neuPrint server. You can install it from PyPI:

pip3 install neuprint-python

navis.interfaces.neuprint wraps neuprint-python and adds a few new functions to fetch and convert data into NAVis objects.

# Import navis
import navis

# Import neuprint wrapper by navis
import navis.interfaces.neuprint as neu

First set up the connection: You can either pass your API token directly or store as NEUPRINT_APPLICATION_CREDENTIALS environment variable. The latter is the recommended way and we will use it here:

client = neu.Client(
    "https://neuprint.janelia.org/",
    # token="MYLONGTOKEN"  # use this to instead pass your token directly
    dataset="hemibrain:v1.2.1",
)

You can use all of neuprint's functions:

mbons, roi_info = neu.fetch_neurons(
    neu.SegmentCriteria(instance=".*MBON.*", regex=True)
)
mbons.head(3)

Out:

/opt/hostedtoolcache/Python/3.10.14/x64/lib/python3.10/site-packages/neuprint/client.py:609: FutureWarning:

Series.__getitem__ treating keys as positions is deprecated. In a future version, integer keys will always be treated as labels (consistent with DataFrame behavior). To access a value by position, use `ser.iloc[pos]`

NAVis has added three functions to neu:

• navis.interfaces.neuprint.fetch_roi: returns a navis.Volume from a ROI
• navis.interfaces.neuprint.fetch_skeletons: returns fully fledged navis.TreeNeurons - nodes, synapses, soma and all
• navis.interfaces.neuprint.fetch_mesh_neuron: returns navis.MeshNeurons - including synapses

Let's start by fetching the mesh for the right mushroom body ROI:

mb = neu.fetch_roi("MB(R)")
mb

Out:

<navis.Volume(name=MB(R), color=(0.85, 0.85, 0.85, 0.2), vertices.shape=(57913, 3), faces.shape=(115856, 3))>

Next, let's fetch the skeletons of all right MBONs:

mbon_skeletons = neu.fetch_skeletons(
    neu.SegmentCriteria(instance=".*MBON.*_R", regex=True), with_synapses=True
)
mbon_skeletons.head()

Co-visualize the MBONs and the MB volume:

navis.plot3d(
    [mbon_skeletons[0], mb],
    legend=False,  # Hide the legend (more space for the plot)
)

Last (but not least), let's make a 2d plot for the tutorial's thumbnail:

import matplotlib.pyplot as plt

fig, ax = navis.plot2d(
    [mbon_skeletons[0], mb],
    c=(0, 0, 0, 1),  # Make the neuron black
    method="3d",
    connectors=True,
    linewidth=0.5,  # Make neuron a bit thinner to emphasize the synapses
    view=("x", "-z"),
)

plt.tight_layout()

All NAVis functions for analysis & visualization should work on these neurons. If not, please open an issue on Github.

Total running time of the script: ( 0 minutes 25.087 seconds)

Download Python source code: plot_00_remote_neuprint.py

Download Jupyter notebook: plot_00_remote_neuprint.ipynb

Gallery generated by mkdocs-gallery