MC-DEPI
Monte-Carlo Diffusion-Enhanced Photon Inference (MC-DEPI) is a method to simulate solution-based single-molecule FRET experiments and to estimate physical parameters to the experimental data.
The method is described in:
Monte-Carlo Diffusion-Enhanced Photon Inference: Distance Distributions and Conformational Dynamics in single-molecule FRET
Antonino Ingargiola, Shimon Weiss, Eitan Lerner (2018) https://doi.org/10.1101/385252
The input is the burst-selected photon data from a smFRET experiment. DEPI simulates photon color and nanotimes, based on donor-acceptor self-diffusion. This allows to accurately simulate smFRET experiment taking into account the FRET-enhancenment due to the donor-acceptor diffusion happening during the fluorescence lifetime.
Notebooks
Example notebooks illustrating how to use this package are available here:
- https://github.com/tritemio/mcdepi2018-paper-analysis
Installation
The most update version of depi can be installed from github:
pip install git+https://github.com/OpenSMFS/depi/ --upgrade
Dependencies
- python >=3.6
- numpy >=1.9
- scipy >=1.0
- matplotlib >=2.0
- cython >= 0.23
- pandas >= 0.22
- joblib >= 0.11