Skip to content

This is the official implementation of our publication "Deep learning enables fast and dense single-molecule localization with high accuracy" (Nature Methods)

License

Notifications You must be signed in to change notification settings

TuragaLab/DECODE

Repository files navigation

DECODE

Gateway Test Unit Tests Docs

DECODE is a Python and Pytorch based deep learning tool for single molecule localization microscopy (SMLM). It has high accuracy for a large range of imaging modalities and conditions. On the public SMLM 2016 software benchmark competition, it outperformed all other fitters on 12 out of 12 data-sets on both detection accuracy and localization error, often by a substantial margin. DECODE enables live-cell SMLM data with reduced light exposure in just 3 seconds and to image microtubules at ultra-high labeling density.

DECODE works by training a DEep COntext DEpendent (DECODE) neural network to detect and localize emitters at sub-pixel resolution. Notably, DECODE also predicts detection and localization uncertainties, which can be used to generate superior super-resolution reconstructions.

Getting started

The easiest way to try out the algorithm is to have a look at the Google Colab Notebooks. We provide them for training our algorithm and fitting experimental data. For installation instructions and further information please refer to our docs. You can find these here:

  • Documentation
  • DECODE Training (NEW: v0.10) Open In Colab
  • DECODE Fitting (NEW: v0.10) Open In Colab

Local Installation

Details about the installation can be found in the documentation.

DECODE cloud

Please reach out to Lucas (lrm@lrm.dev) if you want to use DECODE, but you do not have the right hardware, or want to use it at a larger scale.

Video Tutorial

As part of the virtual I2K 2020 conference we organized a workshop on DECODE. Please find the video below. DECODE is being actively developed, therefore the exact commands might differ from those shown in the video.

DECODE Video Tutorial

Paper

This is the official implementation of the publication.

Artur Speiser*, Lucas-Raphael Müller*, Philipp Hoess, Ulf Matti, Christopher J. Obara, Wesley R. Legant, Anna Kreshuk, Jakob H. Macke†, Jonas Ries†, and Srinivas C. Turaga†, Deep learning enables fast and dense single-molecule localization with high accuracy. Nat Methods (2021). https://doi.org/10.1038/s41592-021-01236-x

Data availability

The data referred to in our paper can be accessed at the following locations:

Contributors

If you want to get in touch, the best way to get your questions answered is our GitHub discussions page

Support

Jakob H. Macke and Artur Speiser were supported by the German Research Foundation (DFG) through Germany’s Excellence Strategy (EXC-Number 2064/1, project no. 390727645) and the German Federal Ministry of Education and Research (BMBF, project no. ADIMEM, FKZ 01IS18052). Srinivas C. Turaga is supported by the Howard Hughes Medical Institute. Jonas Ries, Lucas-Raphael Mueller and Philipp Hoess were supported by the European Molecular Biology Laboratory, the European Research Council (grant no. CoG-724489 to Jonas Ries) and the National Institutes of Health Common Fund 4D Nucleome Program (grant no. U01 EB021223 to Jonas Ries).

Join us

We offer several open positions. Please take a look at the pdf on how to apply.

Acknowledgements

About

This is the official implementation of our publication "Deep learning enables fast and dense single-molecule localization with high accuracy" (Nature Methods)

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published