Open Computational Tools for Freely Diffusing Single-Molecule Fluorescence Analysis

Computation has become a major component of modern science and, as a result, computational reproducibility has become as fundamental as reproducibility in the lab. In this talk, we present a set of open source tools for freely-diffusing single-molecule fluorescence analysis, which were designed with the goal of reproducibility. The common foundation is represented by Photon-HDF5 (www.photon-hdf5.org), a general-purpose file format for timestamp-based fluorescence data, which simplifies data archival and sharing between different analysis software. Using Photon-HDF5 as its primary input format, FRETBursts (http://tritemio.github.io/FRETBursts) is an open source burst analysis software for smFRET data, supporting continuous wave or pulsed excitation in one or two colors. FRETBursts includes time-dependent background estimation, burst search algorithms, burst filtering, population fitting and extensive plotting capabilities. Since a growing number of custom formats can be converted to Photon-HDF5 (through the phconvert utility), FRETBursts can be used to analyze and compare data acquired with a variety of systems. As a second example, PyBroMo (http://tritemio.github.io/FRETBursts), is a 3D Brownian motion simulator for fluorescent particles under confocal excitation. PyBroMo produces smFRET data files in Photon-HDF5 format containing one or multiple FRET populations, uses numerically-computed point spread function and can simulated effects such as out-of-focus particle contribution to background and burst data. PyBroMo data files can be seamlessly analyzed with FRETBursts and or with any other smFRET analysis program supporting Photon-HDF5. Both FRETBursts and PyBroMo, are written in Python, an increasingly popular open language for scientific computing, and use the interactive Jupyter Notebook environment for user-friendly interfacing. Leveraging the strengths on Photon-HDF5, the presented tools form a solid foundation for developing novel analysis methods and for reproducing third-party results. We aim to collaboratively expand this set of tools for single-molecule fluorescence while pursuing the same principles of openness and reproducibility.