Detecting transitions in protein dynamics using a recurrence quantification analysis based bootstrap method

Abstract

BackgroundProteins undergo conformational transitions over different time scales. These transitions are closely intertwined with the protein’s function. Numerous standard techniques such as principal component analysis are used to detect these transitions in molecular dynamics simulations. In this work, we add a new method that has the ability to detect transitions in dynamics based on the recurrences in the dynamical system. It combines bootstrapping and recurrence quantification analysis. We start from the assumption that a protein has a “baseline” recurrence structure over a given period of time. Any statistically significant deviation from this recurrence structure, as inferred from complexity measures provided by recurrence quantification analysis, is considered a transition in the dynamics of the protein.ResultsWe apply this technique to a 132 ns long molecular dynamics simulation of the β-Lactamase Inhibitory Protein BLIP. We are able to detect conformational transitions in the nanosecond range in the recurrence dynamics of the BLIP protein during the simulation. The results compare favorably to those extracted using the principal component analysis technique.ConclusionsThe recurrence quantification analysis based bootstrap technique is able to detect transitions between different dynamics states for a protein over different time scales. It is not limited to linear dynamics regimes, and can be generalized to any time scale. It also has the potential to be used to cluster frames in molecular dynamics trajectories according to the nature of their recurrence dynamics. One shortcoming for this method is the need to have large enough time windows to insure good statistical quality for the recurrence complexity measures needed to detect the transitions.

Understanding the Structure and Apo Dynamics of the Functionally Active JIP1 Fragment

“…To estimate the error associated with the percent occurrences reported in Table , we performed a random sampling with replacement (bootstrapping) analysis; i.e., we randomly sampled our ensemble structure-to-Conformer assignments. , In both cases, the randomly generated sample size was constrained by the original sample size. The population averages from our resampled percent occurrences are in good agreement with our original data (Table ) with percent occurrence errors less than 2%.…”

Non-Linear Analysis of River System Dynamics Using Recurrence Quantification Analysis

“…Moreover, Recurrence Plots and Recurrence Quantification Analysis with epoqs seem to be a useful tool for detecting transitions during the evolution of the system [17]. Since 1989, the RP methodology and also the RQA have been successfully employed in many systems, giving satisfactory results regarding the comprehension of various systems' dynamics during their time evolution, such as biology [18], physiology [19], engineering [20], environmental studies [21][22][23][24], climate change [25][26][27], biology [28], finance [29], biological data [30], chemical processes [31], transportation [32], complex networks [33] and medical data [34], among others.…”

Detecting Transient Trapping from a Single Trajectory: A Structural Approach

“…They can be used to reconstruct protein structure [22] and are even used to detect structural changes in reactiondiffusion systems [23]. In general, they are used for quantifying non-linear time-series derived from dynamical systems, such as detecting protein conformation changes in molecular dynamics [24] or for quantifying physiological measurements [25]. In the context of dynamical systems, it has been shown that one can reconstruct the chaotic attractor associated with a time-series [26].…”