Deciphering anomalous heterogeneous intracellular transport with neural networks

Abstract

Intracellular transport is predominantly heterogeneous in both time and space, exhibiting varying non-Brownian behavior. Characterization of this movement through averaging methods over an ensemble of trajectories or over the course of a single trajectory often fails to capture this heterogeneity. Here, we developed a deep learning feedforward neural network trained on fractional Brownian motion, providing a novel, accurate and efficient method for resolving heterogeneous behavior of intracellular transport in space and time. The neural network requires significantly fewer data points compared to established methods. This enables robust estimation of Hurst exponents for very short time series data, making possible direct, dynamic segmentation and analysis of experimental tracks of rapidly moving cellular structures such as endosomes and lysosomes. By using this analysis, fractional Brownian motion with a stochastic Hurst exponent was used to interpret, for the first time, anomalous intracellular dynamics, revealing unexpected differences in behavior between closely related endocytic organelles.

Local Analysis of Heterogeneous Intracellular Transport: Slow and Fast Moving Endosomes

“…Improved microscopy imaging, tracking and analysis methods revealed the intrinsic spatial and temporal heterogeneity within individual trajectories of numerous biological processes [5,[10][11][12][13][14][15][16][17][18]. Significant progress has also been made in analysis and interpretation of superresolution single particle trajectories [19][20][21][22][23].…”

Variable-order fractional master equation and clustering of particles: non-uniform lysosome distribution

“…However, the exact mechanisms for how lysosomes maintain such a macroscopic spatial distribution remain unclear. The aim of this subsection is to show that lysosomal distributions in the cell can be explained to a large extent by the anomalous mechanism detailed in this paper, since subdiffusion is the most prevalent characteristic in lysosomal movement [21,25]. Anomalous subdiffusion can occur as a result of non-uniform crowdedness [10] in the cytoplasm.…”

“…Quantifying dynamic cellular processes has been a major success for anomalous transport theory and much scientific work is still ongoing [15][16][17][18][19][20]. However, current experimental studies [21][22][23] are finding evidence of heterogeneous anomalous transport in intracellular processes while the theory for heterogeneous anomalous transport (specifically when the fractional exponent μ is no longer a constant) remains largely neglected. In fact, it is given knowledge that the cellular cytoplasm is a vastly heterogeneous complex fluid [24].…”

“…In particular, organelles responsible for cellular metabolism and degradation called lysosomes move predominantly subdiffusively with heterogeneous fractional exponents that depend on their spatial positioning [21,25]. This implies that lysosome dynamics should be adequately described by a fractional diffusion equation with a space-dependent fractional exponent, μ(x).…”

An Estimation Algorithm for General Linear Single Particle Tracking Models with Time-Varying Parameters

“…Sliding (or rolling) windows have also been applied to SPT data, though with the single goal of segmenting the data based on features in the MSD [ 19 , 20 , 21 , 22 ]. More recently, machine learning techniques have been brought to bear on the problem of trajectories with time-varying parameters [ 23 , 24 ]. While results have been promising, there is a need to train the underlying neural networks and as a result there are concerns about transfer learning when applying the methods to different model classes than those used for training.…”