I am a theoretical physicist who is interested in the spatial and temporal organization of biological processes. This organization often emerges from the interplay of many small parts, leading to complex behaviors. I study such problems in cells and tissues using tools from statistical physics, dynamic systems theory, and information theory.

Olfaction

is the sense of smell, which we use to recognize the chemical composition of our surrounding. Despite its importance for all animals, from bacteria to humans, smell is the sense we understand the least. For instance, we do not know how our nose discerns mixtures of the about 10,000 chemicals that we can smell. Another open questions is how different concentrations of smells are distinguished. All this information about an odor must be encoded in the activity pattern of the about 300 different receptor types in a human's nose.

We study the olfactory system by investigating theoretical models of the receptors. We would like to understand how the typically complex odors, consisting of various substrates at different concentrations, are processed. Furthermore, some substrates often appear together in odors, for instance because the originate from the same biological process. Clearly, typical odor mixtures also vary from place to place and from season to season. This structure in typical odor mixtures has likely played a role in evolution to optimize the receptors. By studying simple model as well as naturally occurring receptors, we will uncover design principles of the olfactory system, which can then be used to improved artificial smell sensors.


Publication: Zwicker et al, PNAS (2016) and Zwicker, PLos One (2016)


Schematic picture of the a centrosome
Schematic picture of our model describing the signal processing in the olfactory system. An odor comprised of many ligands excites the olfactory receptors, whose signal is accumulated in respective glomeruli. This first odor representation is transformed using global inhibition mediated by local neurons. The sparse, concentration-invariant second odor representation is subsequently interpreted by the brain.

Fluid dynamics of inhalation and exhalation

All vital tasks of the nose depend critically on the airflow, which is driven by the pressure difference created by the lungs and depends on the complex geometry of the nasal cavity. We want to understand how the geometry influences the flow, so we can learn how to modify it to help treating breathing problems. However, choosing the right treatment also requires considering the other functions of the nose, like humidification, heating, and smelling. The physics of these processes can be conceptualized as the transport of passives tracers by the flow and their exchange with the walls, which are covered in aqueous mucus. We want to understand how the nasal geometry and the absorption properties of the mucus influence these processes. What nasal geometry heats and humidifies the air effectively? Where should the olfactory receptors be placed to effectively sense odorants? Does the odor percept depend on the direction of the airflow? Are humans particularly good at sensing food odors while exhaling? Insights in the physical aspects of olfaction will help to understand the evolution of the nose and to optimize the geometry and surface properties of artificial smell sensors.

Spontaneously dividing active droplets as a model for protocells

The first lifeforms on earth must have been simple enough to emerge spontaneously, but they also needed to divide and propagate, such that natural selection and evolution could lead to more complex lifeforms. So far, the physical properties of such early cells are unclear, but it has been proposed almost a century ago that they could have been liquid-like droplets. We showed that such simple liquid droplets can indeed divide spontaneously if the chemical reaction that builds the droplet material from precursors is driven by an external energy input, like sun light or heat from thermal vents. Here, the chemical reaction of these active droplets plays the role of a prebiotic metabolism.


Publication: Zwicker et al, Nature Physics (2016)
Schematic picture of the a centrosome
Schematic picture of the centrosome comprised of centrioles (blue) and pericentriolar material (orange) organizing microtubules (green). We describe the pericentriolar material as a liquid droplet that organizes around the centrioles.

Centrosomes

are small organelles present in all animal cells. They are important for organizing the mitotic spindle, which segregates the DNA during cell division. Cancer cells often contain more than two centrosomes, which impairs normal cell division. It is thus important to understand the assembly of centrosomes in order to tackle failures in their formation and function.

The centrosome is a dynamic aggregate, which forms and dissolves in synchrony with the cell cycle. Beside reactions between and diffusion of centrosome components, aspects of non-equilibrium thermodynamics have to be considered to describe the observed behavior. In fact, we developed a physical description of centrosomes as autocatalytic droplets, whose formation is controlled by chemical reactions. This project was carried out in close collaboration with the group of Tony Hyman at the MPI of Molecular Cell Biology and Genetics.


Publication: Zwicker et al, PNAS (2014)
See more about the details of the model

Active droplets: Controlling droplets by chemical reactions

Controlling the formation and stabilizing droplets is important in many fields ranging from the food industry to cosmetics and medicine. Furthermore, there is more and more evidence that droplets also play an important role to organize the interior of biological cells. Indeed, we propose that centrosomes are liquid droplets. One problem with liquid droplets is that they try to combine to form on large droplet, which is energetically more favorable.

The video below shows that chemical reactions influencing the physical properties of the droplet material can prevent this droplet coarsening. We study generic physical models of droplet formation under the influence of chemical reactions to identify the necessary conditions where multiple droplets are stable. This improves our understanding of droplet formation inside cells and might also benefit technical applications.


Publication: Zwicker et al, PRE (2015)

Simulation of the coarsening of droplets with (right) and without (left) chemical reactions.
Schematic picture of the core biochemical network of the circadian clock of cyanobacteria
Schematic of the chemical network of the circadian clock

Circadian clocks

keep your body in synchrony with the daily changes in your environment. These clocks are employed ubiquitously by almost all organisms. Even some bacteria have developed a way to tell time in order to prepare themselves during night to harvest light after sunrise. The clock in these simple creatures is surprisingly complex. It consists of two independent oscillators that are intertwined in a complicated manner.

We use stochastic simulations of the involved chemical reactions to elucidate the reason for this apparent redundancy. It turns out that these two mechanisms make the clock robust under all physiological conditions: the one oscillator based on the modification of proteins fails for rapidly growing bacteria. Conversely, the other oscillator based on production and degradation of proteins is very costly for starving bacteria. Consequently, only the combination yields an optimal configuration that can be used as an efficient clock under all conditions. These predictions of the model have recently been confirmed by experiments published in Science.

Publication: Zwicker et al, PNAS (2010)

Automated image analysis

is an important tool in quantitative biology. In many experiments, manual analysis of the movies of modern high-throughput microscopy is a bottleneck. Additionally, using computer algorithms often has the advantage that the bias of the human investigator is reduced and that the reproducibility is increased. Furthermore, a well-composed algorithm can obtain the required data with a higher precision than manual measurements.

In this project, we aimed at analyzing movies of microtubules moving on a functionalized surface. A typical movie would have many of those elongated filaments moving at the same time. Bending and crossing of microtubules makes writing a robust algorithm challenging. We wrote a computer program named FIESTA that analyzes such movies automatically. The program provides an easy to use interface and is able to track both elongated objects and single particles with nanometer precision.

Publication: Ruhnow et al, Biophy. J (2011)