Wavelets analysis is similar to Fourier analysis in that it is a method of examining a signal in frequency space, but, unlike Fourier analysis, wavelet analysis allows one to retain information about when in time specific frequencies are occurring. Additionally, wavelet transforms have the ability to explore different kinds of shapes of motion. Many signals are much better handled by wavelet transforms as a result of these features. One example of this is a signal in which the frequency changes over time. The example below illustrates this difference.
In our simulations, the trajectory of an atom over time is taken as the signal. We did performed wavelet analysis on the Cα atoms of our simulations and found that examining the wavelengths of the wavelet oscillations that occur over time can be an excellent method for quickly locating events of interest in a simulation. This creates a tool both for finding notable pieces of a simulation as well as for quickly screening a large set of simulations for potential differences.