Overview

This project contains the tools to create the data files the BBP Cell Atlas is built on. The creation of atlas files is the first step towards the creation of a circuit.

The tools implementation is based on the methods of A Cell Atlas for the Mouse Brain by Csaba Eroe et al., 2018. The source code was originally written by Csaba Eroe, Dimitri Rodarie, Hugo Dictus, Lu Huanxiang, Wajerowicz Wojciech and Jonathan Lurie.

Atlas building tools operate on data files coming from the Allen Institute for Brain Science (AIBS). These data files were obtained via experiments performed on P56 wild-type mouse brains.

The tools allow to:

• combine AIBS annotation files to reinstate missing mouse brain regions
• combine several AIBS gene marker datasets, to be used as hints for the spatial distribution of glia cells
• split the layer 2/3 of the AIBS mouse isocortex into layer 2 and layer 3
• assign direction vectors or orientations to voxels in a selected brain region
• compute distances between voxels and region boundaries, i.e., the so-called placement hints to be used by the placement-algorithm
• compute cell densities for several cell types including neurons and glia cells in the whole mouse brain
• flatten a laminar brain region by collapsing the streamlines of its fiber tracts direction field, i.e., create a flat map to be used by white-matter-projections

Tools can be used through a command line interface.

Currently, atlas-building-tools is mainly a wrapper, for backwards compatibility, around:

• https://github.com/BlueBrain/atlas-densities
• https://github.com/BlueBrain/atlas-direction-vectors
• https://github.com/BlueBrain/atlas-placement-hints
• https://github.com/BlueBrain/atlas-splitter

Installation

This python project depends on:

• the python package Rtree, which might require the separate installation of the C++ library libspatialindex (see instructions below)
• the BBP C++ toolkits Regiodesics and Ultraliser
• the BBP python-C++ bindings cgal-pybind

The remaining installation instructions are only relevant if you want to install atlas-buidling-tools on BB5 with its latest sources. Prior to running

git clone git@bbpgitlab.epfl.ch:nse/atlas-building-tools.git
cd atlas-building-tools
pip install -i https://bbpteam.epfl.ch/repository/devpi/simple/ -e .

you need to install the dependencies listed below.

Regiodesics

This python project depends on Regiodesics, a BBP C++ toolkit by the Viz Team.

To load Regiodesics on the BB5 cluster, run the following command:

module load regiodesics/0.1.2

For an installation of Regiodesics from the sources, install cmake (version >= 3.5 required), boost and OpenSceneGraph first. The installation of Regiodesics is as follows.

git clone https://bbpgitlab.epfl.ch/nse/archive/regiodesics
cd Regiodesics
git submodule update --init
mkdir build
cd build
cmake ..
make -j
cd ..
export PATH=$PATH:$PWD/build/bin

Ultraliser

This python project depends on Ultraliser, a BBP C++ toolkit by the Viz Team.

To load Ultraliser on the BB5 cluster, run the following command:

module load unstable ultraliser/0.2.0

For an installation of Ultraliser from the sources, install cmake (version >= 3.5 required) and proceed as follows.

git clone https://github.com/BlueBrain/ultraliser
cd ultraliser
mkdir build
cd build
cmake ..
make -j
cd ..
export PATH=$PATH:$PWD/build/bin

Rtree

This python project depends on Rtree, a python package which requires the libspatialindex library, a C++ dependency.

If you are using conda, then libspatialindex should be installed automatically with Rtree.

If this is not the case, you can install libspatialindex via brew on MacOS or via apt-get on Ubuntu systems.

On the BB5 cluster, install rtree and its dependency libspatialindex with:

module load unstable py-rtree/0.8.3

cgal-pybind

The BBP python project cgal-pybind contains python bindings for several functions of the CGAL C++ library. The algorithm of atlas-building-tools which creates a flat map uses specifically CGAL's authalic map.

On the BB5 cluster, install cgal-pybind with:

module load unstable py-cgal-pybind/0.0.2

poisson-recon-pybind

The BBP python project poisson-recon-pybind contains python bindings for the reconstruction surface algorithm of PoissonRecon.

On the BB5 cluster, install poisson-recon-pybind with:

module load unstable py-poisson-recon-pybind/0.1.0

Instructions for developers

Run the following commands before submitting your code for review:

cd atlas-building-tools
isort -l 100 --profile black atlas_building_tools tests setup.py
black -l 100 atlas_building_tools tests setup.py

These formatting operations will help you pass the linting check testenv:lint defined in tox.ini.

Acknowledgements

The development of this software was supported by funding to the Blue Brain Project, a research center of the École polytechnique fédérale de Lausanne (EPFL), from the Swiss government’s ETH Board of the Swiss Federal Institutes of Technology.

For license see LICENSE.txt.

Copyright © 2020-2024 Blue Brain Project/EPFL