data simulator

The matlab script used to simulate synthetic smFRET trajectories since July 2019 can be found here.

(The earlier version remains available here.)

The focus of the simulation is on empirically reproducing typical smFRET data. To this end, we evaluated experimental smFRET trajectories (from different detectors, dyes, biomolecules) regarding intensities & intensity distributions, noise characteristics, trace lengths etc. Based on this, we obtained parameters that quantitatively describe typical smFRET trajectories.

Using the experiment-derived parameters, each state trajectory is simulated in continuous time according to the supplied transition rate matrix. Next, the traces are discretised using a typical sampling rate of a camera-based smFRET setup. The intensity at time t is obtained by integration over the populations of each state occurring during the corresponding exposure time (thus yielding intermediate intensity levels if a state transition occurred during the exposure time). Optionally, degenerate states are introduced.

The following parameters are used to simulate real-world experiments:

• the transition rate matrix
• frame rate and bleach rate
• minimal and maximal trace length
• the Gaussian* distribution of the fluorescence intensity of each state (leading to not necessarily Gaussian FRET efficiency distributions)
• the number of traces
• possible blinking kinetics of the dyes
• two parameters controlling a possible per-trace variability of the intensity levels

* Although strictly speaking Poissonian, the photon counts observed in typical camera-based smFRET experiments (N>10) are very well modeled by Gaussian distributions.