Project
EF4-4
Diffusion in Dynamically Crowded Cells: Stochastic Models from Spatiotemporal Motion Data
Project Heads
Felix Höfling, Carsten Hartmann
Project Members
Arthur Straube (FU, from 07/19 to 12/19), Upanshu Sharma (FU, from 10/19 to 12/20)
Project Duration
01.01.2019 – 31.12.2021
Located at
FU Berlin
Description
Diffusion in cellular environments, consisting of a variety of interacting entities, is a multiscale process. The project’s objective is to infer effective stochastic models and to quantify memory, using novel stochastic modelling and data assimilation techniques, based on data from experiments and simulations.
Diffusing diffusivity (DD) models
Macromolecular crowding is characterised by a dense and heterogeneous packing of differently sized, mobile and immobile components. From the interaction between these entitites, complex behaviour can emerge on large scales, including long time scales and persistent memory. Experiments on labelled molecules frequently show non-Markovian or non-Gaussian behaviour or both, depending on the window of time scales probed [1].
A description by an idealised Brownian motion is, at best, approximate, and may lead to significant modelling errors with respect to, e.g., spatial particle distributions and their reaction kinetics. Accurate and efficient modelling demands linking complex spatio-temporal data to suitable stochastic models, which should be able to accomodate the feature-rich experimental observations, yet allow for the robust estimation of parameters.
Non-Gaussian transport, yet with a linearly growing mean-square displacement (MSD) has attracted recent interest, whereas previous research almost exclusively focussed on the striking observation of (transient) subdiffusion [2]. A stochastic model that partially captures the former treats the particle’s effective diffusivity as a stochastic mean-reversion process, justified by the constantly rearranging environment [3]. In the most recent version of this “diffusing diffusivity (DD)” model, the molecule’s displacement 𝑋 obeys a non-degenerate diffusion that bears some resemblance with the well-known Cox-Ingersoll-Ross (CIR) process in mathematical finance. The full DD model is Markovian. However, the dynamics of 𝑋 alone may display significant memory effects or non-Markovianity, unless the system exhibits a clear time scale separation.
Parameter estimation of multiscale systems
We will derive a hierarchy of the effective models for the diffusing particle in various scaling regimes and study properties of the parameter estimator for the effective diffusivities. When multiple scales are present, statistical estimators, such as the maximum likelihood estimator, are known to suffer from a systematic bias unless clever subsampling strategies are employed, and one part of the project consists in studying the properties of non-parameteric and semi-parametric estimators for the diffusivities. For finite time scale separation we expect, however, an effective model to exhibit significant sub-exponential correlations, which suggests to systematically seek non-Gaussian approximations, akin to the (non-Markovian) Kac-Zwanzig heat bath model.
Capturing non-Markovianity and memory effects
One possibility to describe non-Markovian effects is by means of the generalised Langevin equation (GLE) that, in its most frequently used Gaussian form, is specified by the autocorrelation function (ACF) of the noise increments or, equivalently, its memory kernel. A model-free alternative studied in this project uses higher-order memory functions that allow for the consistent interpolation from short to long time scales. The approach is based on a Fourier representation of the ACF of a stationary process in terms of a spectral density (e.g. [1]). The key quantity here, which is experimentally accessible, is the characteristic function of the process (i.e. the Fourier transform of its spectral density) which after systematic short-time approximations captures non-Markovian processes with persistent memory (long-time anomalies) as well as non-Gaussian transport with spatiotemporal memory or non-Gaussian noise, beyond the standard GLE.
A scientific novelty is that a completely data-driven framework is developed to study memory and crowding for diffusing particles that incorporates many available standard approaches, such as the GLE framework, and that rests on very few generic assumptions. The project explores the potential and limitations of the class of DD models, based on results of complex analysis and on actual data. The analysis of ACF including the frequency dependence is still in its infancies. Estimating spatiotemporal memory from trajectory data is new and goes beyond standard GLE approaches.
References
[1] F. Höfling and T. Franosch. Anomalous transport in the crowded world of biological cells. Rep. Prog. Phys., 76:046602, 2013.
[2] A. V. Chechkin, F. Seno, R. Metzler, and I. M. Sokolov. Brownian yet non-gaussian diffusion: From superstatistics to subordination of diffusing diffusivities. Phys. Rev. X, 7:021002, 2017.
[3] Y. Lanoiselée and D. S. Grebenkov. A model of non-Gaussian diffusion in heterogeneous media. J. Phys. A: Math. Theor., 51:145602, 2018.
Project Webpages
Selected Publications
- C. Hartmann, L. Neureither, and U. Sharma,
Coarse-graining of non-reversible stochastic differential equations: quantitative results and connections to averaging,
SIAM J. Math. Anal., 52(3):2689–2733, 2020 (preprint). - A. V. Straube, B. G. Kowalik, R. R. Netz, and F. Höfling,
Rapid onset of molecular friction in liquids bridging between the atomistic and hydrodynamic pictures,
Commun. Phys., 3:126, 2020. - D. Frömberg and F. Höfling,
Generalized master equation for first-passage problems in partitioned spaces,
J. Phys. A: Math. Theor., 54:215601, 2021 (preprint). - Z. Mezdoud, C. Hartmann, M. R. Remita, and O. Kebiri.
α-hypergeometric uncertain volatility models and their connection to 2BSDEs,
Bull. Inst. Math., Acad. Sin., 16(3):263–288, 2021 (preprint). - H. Bouanani, C. Hartmann, and O. Kebiri,
Model reduction and uncertainty quantification of multiscale diffusions with parameter uncertainties using nonlinear expectations,
submitted manuscript, arXiv:2102.04908, 2021. - T. Breiten, C. Hartmann, L. Neureither, and U. Sharma,
Stochastic gradient descent and fast relaxation to thermodynamic equilibrium: a stochastic control approach,
J. Math. Phys. 62(12):123302, 2021 (preprint). - U. Sharma and W. Zhang,
Non-reversible sampling schemes on submanifolds.
SIAM J. Numer. Anal., 59(6):2989–3031, 2021 (preprint). - F. Legoll, T. Lelièvre, and U. Sharma,
An adaptive parareal algorithm: Application to the simulation of molecular dynamics trajectories,
SIAM J. Sci. Comput., 44(1):B146–B176, 2022 (preprint). - R. I. A. Patterson, D. R. M. Renger, and U. Sharma.
Variational structures beyond gradient flows: a macroscopic fluctuation-theory perspective,
submitted manuscript, arXiv:2103.14384, 2021.
Selected Pictures
Anomalous diffusion of a probe particle in a crowded medium.
Fluorescence correlation spectroscopy (FCS) reveals anomalous, yet Gaussian diffusion in a crowded medium. For more information, see Banks et al., Soft Matter, 12:4190, 2016.
Smooth yet chaotic motion of atoms in a liquid (Journal cover art)
Spectral analysis of memory shows that the smooth yet chaotic motion of atoms in a liquid leads to a rapid onset of friction below a characteristic frequency. More information can be found in Straube et al., Commun. Phys., 3:126, 2020.
Please insert any kind of pictures (photos, diagramms, simulations, graphics) related to the project in the above right field (Image with Text), by choosing the green plus image on top of the text editor. (You will be directed to the media library where you can add new files.)
(We need pictures for a lot of purposes in different contexts, like posters, scientific reports, flyers, website,…
Please upload pictures that might be just nice to look at, illustrate, explain or summarize your work.)
As Title in the above form please add a copyright.
And please give a short description of the picture and the context in the above textbox.
Don’t forget to press the “Save changes” button at the bottom of the box.
If you want to add more pictures, please use the “clone”-button at the right top of the above grey box.