Superresolution visualization of 3D protein localization data from a range of microscopes

Read, analyze and visualize SMLM localization data quickly from a wide range of localization algorithms.

Example rendering of EPFL challenge dataset 'MT0.N1.HD'

See demo.ipynb for example usage.

Requirements

Python

• vtk
• numpy
• jupyter
• pandas
• seaborn
• requests
• scipy

See requirements.yml for Conda, piprequirements.txt for pip

Optional

• Paraview

Gif of install

Supported Microscopes/Algorithms

• EPFL Challenge:
  • epflreader.EPFLReader('data.csv')
  • source
• Leica GSD
  • gsd.GSDReader('test.bin') # with test.desc in same folder,
  • gsdreader.GSDReader('test.ascii', preprocess=True, binary=False) # ASCII format
  • (needs pixel to nm conversion (e.g. *160 nm /px) : obj.points *= X
  • source
• Tafteh et al dSTORM with z-drift correction (LSI - UBC)
  • dlpreader.DlpReader('test.3dlp')
  • source
• Rainstorm
  • db = rainstormreader.RainStormReader('data.csv') # Automatically finds pixel to nm
  • source
• Abbelight
  • ab = abbelightreader.AbbelightReader('data.csv') # in Nm
  • source

Install

PIP

$pip install smlmvis

Conda

$conda conda install -c bcardoen smlmvis 

Local from git master

$git clone git@github.com:bencardoen/smlmvis.git
$pip install .

Optional

You may want to install optional dependencies, e.g. jupyter notebook and seaborn for the demo:

pip install jupyter seaborn

Usage

See demo.ipynb for example usage.

A typical workflow is

• use one of the readers (e.g. GSDReader in smlmvis.gsreader) to load in the SMLM data
• process the point cloud (obj.points) or compute statistics on the metadata (obj.values)
• write out the data to vtk/paraview format using e.g. VtuWriter in vtuwriter

Cite

@misc{Cardoen2019,
  author = {Cardoen, Ben},
  title = {Superresolution visualization of 3D protein localization data from a range of microscopes},
  year = {2019},
  publisher = {GitHub},
  journal = {GitHub repository},
  doi = 189660035
  howpublished = {\url{https://github.com/bencardoen/smlmvis/}}
}

Tests

See tests/test_writer.py

This will download the challenge data set, read it, decode it, write it to VTK and compare with a reference.

Acknowledgements

VTU writing code uses the VTK examples heavily to figure out how to interface with VTK.

• VTK Python API