fair-software.nl recommendations	Badges
1. Code repository
2. License

STAPL-3D (SegmenTation Analysis by ParaLlelization of 3D Datasets) is an analysis tool for mLSR-3D or other high dimensional 3D fluorescency microscopy datasets.

In brief, the STAPL-3D pipeline is characterized by:

• distributed processing with deployment on computing clusters or workstations;
• inclusion of quality assurance reports;
• modularity of processing steps;
• maximal data retrieval;
• focus on spatial aspects of the data.

What STAPL-3D does to allow accurate analysis can be summarized as:

• correction of inhomogeneity artefacts;
• segmentation with subcellular resolution;
• zipping of blocks for lossless analysis;
• feature extraction for spatio-phenotypic patterning;
• backprojection for 3D exploration of analysis results in native imaging space.

STAPL-3D is created for deployment on High Performance Computing clusters and tested with SGE and SLURM job schedulers. STAPL-3D can also be run on (high-end) workstations. Our lab uses HP Z8 G4 with dual Intel Xeon Gold 5122 3.6 GHz processors, 1 TB RAM, 3x2TB SSD + 6x5TB HDD in RAID-5, and an Nvidia Quadro P4000 8GB graphics card. This workstation setup can process a 100 x 15000 x 15000 x 8 voxel dataset within a day. More modest systems are also feasible: the RAM requirements scale with the chosen parallelization blocksize and is approximately 40GB for a blocksize of 100x1280x1280 voxels for the membrane enhancement step. We would recommend a minimum of 16GB of RAM.

This software runs on Linux, MacOS and Windows. It has been tested on:

• Linux: CentOS7
• MacOS: Big Sur 11.6.1
• Windows: 10 Pro

dependencies: Python (auto-installed through with conda or pip)

• python>=3.8
• scikit-image
• h5py
• matplotlib
• numpy
• pyyaml
• scipy
• pandas
• pypdf2
• czifile
• nibabel
• simpleitk

Other software packages:

• ACME
• Ilastik (optional)
• Fiji (optional)

STAPL-3D is most easily installed (< 5 min) with conda by pasting the following code in the terminal:

git clone https://github.com/RiosGroup/STAPL3D.git
cd STAPL3D

conda env create -f environment.yml
conda activate stapl3d
pip install .

To use membrane enhancement in the pipeline, ACME needs to be installed. To use the machine-learning channel clean-up procedures, Ilastik needs to be installed. To run the demos, these two external packages are not required as we provide these steps.

Jupyter notebooks are provided in the demos directory to showcase the basic functionality of STAPL-3D. The data for the demo (~6GB) can be downloaded here.

• preprocessing_demo (runtime: 20 min)
• segmentation_demo (runtime: 20 min)
• feature_extraction_demo (runtime: 10 min)

If you installed STAPL-3D using conda, you can start a notebook in the stapl3d enviroment as follows:

conda activate stapl3d
python -m ipykernel install --user --name=stapl3d
jupyter notebook

• copy the file stapl3d/pipelines/.stapl3d.ini to your HPC home directory

• adapt the paths in .stapl3d.ini for:

  • STAPL3D: stapl3d package directory
  • ACME: directory with the ACME binaries
  • FIJI: path to fiji executable
  • ILASTIK: path to run_ilastik.sh

• create a directory <datadir> for the <dataset>
• generate a parameter file <datadir>/<dataset>.yml for your dataset: use stapl3d/pipelines/params.yml as a template
• upload the datafile <datadir>/<dataset>.czi
• an example pipeline for HPC usage is provide in stapl3d/pipelines/pipeline.sh
• an example pipeline for python usage is provided in stapl3d/pipelines/pipeline.py

If you want to contribute to the development of STAPL3D, have a look at the contribution guidelines.

Copyright (c) 2020,

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

This package was created with Cookiecutter and the NLeSC/python-template.