DTU ACTIVE : 14 PhD Projects
I. Molecular Level
PhD1. Cost-accuracy trade-offs in active systems (M. Esposito, T. Schmidt, A. Tkatchenko)
At equilibrium, the probability for an open system to be in a given state is a decreasing exponential of the energy of the state and all currents (e.g. charge or energy) through the system are zero. By driving the system away from equilibrium, one can focus probabilities on specific degrees of freedom which can be collective and very delocalized and one can produce currents. We want to characterize trade-offs between the accuracy with which probabilities and currents can be controlled and the energy cost to achieve it. Progress has been achieved in recent years with the discovery of thermodynamic uncertainty relations (i.e. relations which provide universal bounds relating precision of given observables to the dissipation needed to produce it). However, the bounds obtained are often too weak. We gathered evidence that by further specifying the class of observables that one considers or by considering specific models, these bounds can be dramatically improved. We plan to apply these concepts in the context of molecular computing (this field is concerned with performing logical operations using chemical reaction networks such as DNA strand displacement ones. The long term goal of this project is to understand how subcellular processes can reach high reproducibility by operating far from equilibrium.
PhD1–supervised by M. Esposito in collaboration with T. Schmidt, A. Tkatchenko, I. Poltavsky, M. Polettini.
PhD2. Nonequilibrium quantum-mechanical dispersive forces (A. Tkatchenko, M. Esposito, T. Schmidt)
A priori, there is no unambiguous way to define a force between two active particles. Let us consider two underdamped particles interacting via a conservative potential where particle 2 is driven by a nonconservative force only acting on 2 and one coarse-grains particle 2 so that we are left only with a dynamics for particle 1. In this case the effect of hidden particle 2 on 1 is to produce a force acting on 1 that is generically nonconservative (i.e. velocity dependent) [Herpich, Shayanfard, Esposito, soon on arxiv]. We want to understand under what special limits such a force can give rise to a conservative one. We will consider several models aiming to achieve qualitative understanding and quantitative modeling of molecular interactions out of equilibrium. We will start with model geometries (chains, rings, and layers) under applied thermal and chemical gradients, optical excitations, and more general nonconservative driving forces. We will model interactions in the system with an explicitly quantummechanical many-body dispersion (MBD) Hamiltonian. This will allow us to understand the interplay between external driving forces and internal intermolecular interactions in biological systems. We will investigate the fundamental variables that can be defined when external non-conservative forces strongly interact with intermolecular forces within a given system and look at order/disorder transitions in active systems. Our final aim is to study a model of relevance to metabolic reactions, where local chemical reactions influences non-covalent molecular interactions (folding, enzymatic action, etc). An initial application of the MBD framework to protein folding in explicit water already suggests that nonequilibrium driving of plasmonic UV modes could allow controlling folding/unfolding events.
PhD2–supervised by A. Tkatchenko with M. Esposito, T. Schmidt.
PhD3. Microscopic model systems for active quantum matter (T. Schmidt, E. Fodor, M. Esposito)
Brownian motion is an important example for the random motion of a particle in contact with a thermal environment. Its quantum mechanical counterpart, quantum Brownian motion, is an extremely important concept for the description of many phenomena observed in quantum systems in a thermal bath. The understanding of quantum Brownian motion has been greatly advanced by the discovery of solvable microscopic models which give rise to the expected dynamics on a macroscopic scale, e.g., the Caldeira-Leggett model and the spin-boson model. At this macroscopic level, active particles differ from (passive) Brownian particles by the existence of a self-propulsion term in the effective Langevin equation describing their motion. In this project, we would like to develop simple microscopic models systems which result in active motion. In recent years, several efforts have been undertaken to understand active motion at a microscopic scale: it was realized that spin-orbit coupling can be a useful paradigm for understanding active matter at a microscopic scale. Moreover, it was shown that low-dimensional spin systems can model flocking behavior of active particles. The necessary ingredients for such microscopic models are thus a mechanism for bringing the system out of equilibrium, such as an external drive, and a mechanism for translating such a perturbation into a directed motion, such as spin-orbit coupling. In this project, we will thus develop such microscopic fully quantum mechanical motion which give rise to active particles at a macroscopic scale. For this, we will use our group’s previous experience in the modelling of quantum spin systems, spin-orbit coupling, hydrodynamic flow in electron systems, and the study of correlated phases in quantum systems.
PhD3–supervised by T. Schmidt with E. Fodor, M. Esposito.
PhD4. Protein-protein interactions (A. Tkatchenko, A. Skupin, P. May)
Cellular functionality is based to a large extent on proteins and their interactions. While “passive” proteins such as tubulin contribute often to structural properties of cells, “active” proteins known as enzymes are essential for a plethora of biochemical processes which keep cells away from the equilibrium and allow for adaptation. Mutations in protein coding genes and resulting modifications of the protein structure can lead to significant impairment of the protein stability and function and are often associated with diseases. In particular, mutations associated with enzymes of the central carbon metabolism are related to metabolic diseases and to neurodegeneration. Despite the advancements during the last decades, our understanding of the functional consequences of mutations is still rather limited and how quantum-mechanical effects may contribute is barely understood. Here we propose to focus on the chaperon DJ-1 which has been shown to be essential for mitochondrial metabolism and mutations of which are associated with Parkinson’s disease. Based on our preliminary data obtained by molecular dynamics of the DJ-1 dimer using Gromacs, we will investigate the effect of long-range van der Waals forces and predict how these will affect metabolic activity. The theoretical predictions will be validated by enzymatic essays and in vitro experiments using patient derived induced pluripotent stem cells (see PhD 12) and a first demonstration of long-range van der Waals effects on enzymatic activity.
PhD4–supervised by A. Tkatchenko with J. Berryman, A. Skupin, P. May.
PhD5. Membrane-biomolecule interactions (S. Bordas, A. Tkatchenko)
The function of biomolecular systems often depends on binding/unbinding events with different surfaces, such as cell membranes. Such adhesive phenomena depend on the interplay of classical electrostatics, quantum van der Waals dispersion interactions, and mechanical deformation of soft membranes. State of- the-art models for adhesion employ phenomenological treatment of mechanical deformations and electrostatic effects. Here, we propose to go significantly beyond phenomenological models by deriving coarse-grained constitutive relations directly from atomistic quantum-mechanical simulations of biomolecules interacting with membranes. As a proof of principle, we have already demonstrated the potential of direct coupling of quantum and continuum mechanics on the example of delamination of graphene from Silicon surfaces (P. Hauseux et al., Nature Communications, in review). Remarkably, such direct coupling of quantum atomistic models with continuum models of traction-separation law allows us to observe ultra long-range adhesion effects that extend up to 100s of nanometers between graphene and the underlying surface. Within the scope of this project, the PhD student will derive a mechanical force field model of biomolecule and membrane, parametrize the quantum MBD Hamiltonian for both systems, and then simulate the interactions and vibrations of coupled molecule/membrane geometries, aiming to derive constitutive traction-separation laws (TSL) for soft matter. The outcome will be a set of models and a dataset of TSLs for a wide range of biomolecular systems (proteins, DNA, RNA) with membrane surfaces.
PhD5–supervised by S. Bordas with A Tkatchenko, I. Poltavsky.
II. Cellular Level
PhD6 & PhD7. Energy transduction in metabolism and cost of metabolite repair (M. Esposito, C. Linster)
It was recently understood that intracellular repair mechanisms operate not only at DNA or protein level, but also at the level of metabolites, i.e. small molecules. Contrary to what is commonly accepted, metabolic enzymes are not perfect and may catalyze side reactions which can produce toxic side products. Dedicated repair enzymes have been found to keep the level of these toxic side-products very low. While qualitatively understood, quantitative modelling of these mechanisms has never been undertaken. Furthermore, the energetic cost for operating these repair mechanisms has not been analyzed. Our aim will be to do so using our expertise in thermodynamics of chemical reaction networks. Some methodological analogies may exist with studies on the cost of kinetic proofreading using stochastic thermodynamics. Experiments will be performed to first establish and then test the models. On the molecular level, this will involve the production and purification of enzymes of interest and the set-up of enzymatic assays to characterise the kinetic properties of these enzymes. In addition, yeast cells without or with metabolite repair deficiencies will be cultivated in controlled conditions in bioreactors allowing for real-time monitoring of a number of parameters (e.g. oxygen consumption, pH). Here, time-course endo- and exometabolomics measurements will be coupled with metabolic modeling efforts to start addressing the question of the metabolic cost of metabolite repair at the cellular level.
PhD6–supervised by M. Esposito with C. Linster, on the theory and modelling of metabolite repair.
PhD7–supervised by C. Linster with M. Esposito, on the experiments.
PhD8 & PhD9. Cost of Chemical Signaling (A. Skupin, M. Esposito, C. Linster, S. Bordas)
Cells must signal in order to adapt to environmental changes. They typically detect extracellular signals by receptors on the cellular plasma membrane. These signals are transferred to the cellular downstream
processes by second messengers which can trigger a variety of cellular responses. To enable this information transfer and information processing, cells require energy which is predominantly provided by mitochondria in eukaryotic cells. Thus, the crosstalk between cellular signaling and energy metabolism is essential for proper cellular function. A central mechanism of this adaptation is the crosstalk between the versatile second messenger calcium and mitochondrial activity where we have currently shown that the stochastic character of calcium spiking can have a significant effect on energy production and calcium overload in mitochondria can significantly contribute to neurodegeneration in models of Parkinson’s disease. Based on our established modelling framework and experimental data, we hypothesize that the crosstalk between calcium signaling and energy metabolism is of central importance for cell differentiation and fate. Therefore, we propose to develop a firm non-equilibrium description of calcium induced metabolism and test the resulting hypotheses on medically relevant model systems for induced pluripotent stem cell differentiation (with PhD 12) and brain tumor derived cell lines.
PhD8–supervised by A. Skupin with M. Esposito, S. Bordas, on developing a stochastic thermodynamic framework for central energy metabolism and its regulation (including DJ-1 and mitochondrial activity).
PhD9–supervised by C. Linster with A. Skupin, on experiments using our established perfusion system to characterize signaling induced metabolic profiles and mutation specific activation patterns.
III. Population Level
PhD10 & PhD11. Complex behaviors (phase transitions) from activity and self-propulsion (A. Sengupta, E. Fodor, A. Skupin)
Understanding how microbes interface, exchange and communicate with their local surroundings is central to the grand quest for a theory of microbial ecology. Currently we lack a biophysical framework that could explain, generalize, and crucially, predict the if-s, the how-s, and the why-s of the microbe-environment interactions. Research under this theme aims to fill the gap by interfacing active matter physics and fluid mechanics with microbiology and genetic engineering, with a central hypothesis that microbes, across individual, species and community scales, are inherently coupled to their microenvironments, and that their behavioural and physiological traits emerge as a consequence of active biophysical feedbacks between the material, information and energy transport processes. The generality of the results will be demonstrated via trait-based modelling (AnS-EF), specifically in the context of fluctuations in microbial environments representing diverse ecological and medical settings (AnS-AlS). Results from this research will be central to deciphering the biophysical principles underpinning emergence of microbial traits that determine microbial fitness, succession, and selection (see PhD 10), and in developing next generation of databacked bottom-up theory (see PhD 11), not least for their emerging prospects in medical diagnostics, biotechnology, and bioremediation during current climatic trends.
PhD10–supervised by A. Sengupta, on emergent microbial traits under fluctuating environments, using experiments with prokaryotic and eukaryotic single-celled microbes.
PhD11–supervised by A. Sengupta with A. Skupin, E. Fodor, C. Linster, M. Esposito, on bet-hedging strategy and microbial fitness analysis.
PhD12. Energy metabolism induced cell differentiation and cellular interactions (A. Skupin, C. Linster, J. Goncalves, S. Bordas)
While increasing evidence points to the central role that energy metabolism may play in development and neurodegeneration such as Parkinson’s Disease (PD) and Alzheimer’s Disease (AD), our
understanding of these processes is still rather limited because systematic omics approaches at single cell resolution were not available. Based on our current achievements, we propose to investigate how a modified energy metabolism is affecting differentiation. For this purpose, we will differentiate human derived iPS cells with and without mutations in the chaperone DJ-1 protein which is known to be involved in mitochondrial activity in dopaminergic neurons and is linked to Parkinson’s disease. Applying different metabolic inhibitors to change the balance between glycolysis and mitochondrial metabolism followed by single cell RNAseq, we will generate high dimensional single cell data to disentangle the crosstalk between metabolism and cell fate. The resulting data will be also integrated within PhDs 4, 8 and 14.
PhD12–supervised by A. Skupin with C. Linster, J. Goncalves, S. Bordas.
Connections between levels
From level I to Level II
PhD13. Towards field theories of active systems (E. Fodor, M. Esposito, T. Schmidt)
To capture the emerging behavior of active systems, different classes of hydrodynamic theories have been proposed. Some are based on symmetry arguments, determined mostly by the properties of microscopic constituents of the systems. Others are derived via systematic coarse-graining methods starting from the underlying particle-based dynamics. In the latter, phenomenological closures are employed to obtain a closed set of equations for the hydrodynamic variables, such as density and polarization fields. An open question is to understand how these closures affect the thermodynamics of hydrodynamic theories: how does the energetics at micro vs. hydro levels differ? And then, can we provide refined closures which maintain the connection between these different levels? Inspired by recent progress on coarse-graining chemical reaction networks, this project ambitions to propose novel field theories to study the thermodynamics of active matter, with broad applications ranging from bacterial population dynamics and collective behavior of synthetic self-propelled agents.
PhD13–supervised by E. Fodor with M. Esposito, T. Schmidt.
From level II to level III
PhD14. Differentiation as epigenetic turbulences (J. Goncalves, A. Skupin, E. Fodor, S. Bordas)
Recent single cell sequencing technologies have revolutionized our perspective on cellular heterogeneity and cell subtypes. While the resulting data is now intensively used for population characterization and the identification of intracellular mechanisms, a solid framework to investigate intercellular communication that is shaping the population from this kind of data is still lacking. To address this challenge, we propose to combine our machine learning based approaches with mechanistic modelling to understand the underlying principles of cellular heterogeneity and multicellularity. In particular, we will make use of the distributions over large cell populations and apply different approaches to achieve high efficient, scalable and high precision models. In particular, we will consider master equations and their approximations such as Langevin, Fokker-Planck equations, active Brownian particles (ABP) and pseudo time. Moreover, we will explore tradeoffs between different levels of model class complexity such as linear and nonlinear (parametric or basis functions), and different levels of stochastic details. The resulting data and methodologies will give new insights into the interplay between energy metabolism and cell fate by generating a mechanistic understanding on the basic requirements of intra- and intercellular interactions to generate a robust epigenetic landscape by “epigenetic turbulences” as a self-organization mechanism. The developed framework will further integrate results of PhDs 9 and 12 and be linked to PhD 8.
PhD14–supervised by J. Goncalves with A. Skupin, E. Fodor, S. Bordas.