Fermenting bacteria and archaea developed specific enzyme complexes for energy conservation that completely differ from those of the respiratory chain. Glutaconyl-CoA decarboxylase (Gcd), a family member of Na⁺-translocating biotin-containing decarboxylases couples the exergonic decarboxylation of glutaconyl-CoA to crotonyl-CoA with the formation of an electrochemical Na⁺-gradient across the membrane. According to the current mechanistic proposal the carboxylate of glutaconyl-CoA is first transferred onto biotin implicating a large conformational change. Subsequently, the decarboxylation of carboxybiotin drives Na⁺translocation, which was demonstrated for all Na⁺-translocating biotin-containing decarboxylases using inverted membrane vesicles or purified enzymes after incorporation into artificial liposomes. 

The oligomeric composition of the Gcd complex and therefore also the molecular mass are unknown. For molecular and low-resolution structural studies we initiated native gel, multi-angle light scattering, Laser-induced liquid bead ion desorption mass spectrometric and preliminary electron microscopic experiments.

Cortical activity patterns in primates and carnivores typically comprise domains of coactive neurons with an approximately regular spacing on the order of 1mm. Such modular organization is seen in spontaneous activity early in the developing cortex and it persists even after deactivation of LGN suggesting a cortical origin [Smith et al., Nat Neurosci 2018]. However, a biologically plausible and robust mechanism for the intra-cortical generation of modular activity is still missing. Previous network models generating such activity patterns either assume short-range excitation and longer-ranging inhibition (Mexican hat connectivity) [Ermentrout, ROPP 1998] or rely on fast inhibition [Pinto & Ermentrout, SIAM 2001; Kang et al., PNAS 2003], but these assumptions appear in conflict with current experimental evidence. Here we show that spatially localized self-inhibition, consistent with the observed prevalent autapses of fast-spiking interneurons in cat visual cortex [Tamas et al., J Neurosci 1997], relaxes the constraints in the above models, such that biologically more plausible network motifs with shorter ranging inhibition than excitation robustly generate modular activity. Considering a network of excitatory and inhibitory rate units we use linear stability analysis to determine the boundaries of the parameter regime for the formation of modular activity and show that the size of the regime consistent with experimental evidence increases with the strength of local inhibition. A critical prediction of our model is the decrease in spacing of active domains when the total amount of inhibition increases. Indeed, consistent with an increase of inhibition in the developing visual cortex [White & Fitzpatrick, Neuron 2007], we observe that the spacing of spontaneously active domains decreases with age in early ferret visual cortex. We conclude that cortical recurrent networks involving local self-inhibition can robustly generate modular activity patterns.

Heavy-ion collisions are a well-established tool to study the properties of the matter surrounding us. In fact, the state of our universe shortly after the Big Bang can be reproduced by means of colliding two nuclei at velocities close to the speed of light. The energies in such collisions are high enough to destroy the nuclei and create a so-called quark-gluon plasma. In this plasma, the smallest building blocks of matter, quarks and gluons, can move freely and are not bound into protons and neutrons. As the system evolves, the plasma cools and expands and quarks and gluons eventually recombine to form bound states. Consequently, free quarks and gluons cannot be observed under usual conditions. Up to now, heavy-ion collisions remain the only tool to access quarks and gluons in order to study their properties and also the properties of matter. 

In this talk, an introduction to the field of heavy-ion physics is presented. Background information in terms of the standard model of particle physics and the strong interaction is provided. Accelerator facilities, where heavy-ion collisions are carried out, are further introduced. Finally, the different stages of heavy-ion collisions, and how they can be described by theorists, are discussed in detail. 

A model is the representation of an idea, a system that describes phenomena that cannot be experienced directly; models link theory with experiments and allow predictions to improve experimental set ups. Due to technological progress (e. g. in microscopy) biology shifts more and more into a quantitative science. A lot of data in vitro as well as in vivo is accessible for different systems.

A multicellular system is built bottom up as a result of decisions, behaviours, and interactions of individual cells. Mathematical biology offers a powerful tool to understand how individual cells contribute to the formation of a distinct structure. In this talk, different approaches to cell-based modelling are presented (i. e. lattice-based and off-lattice-based models).

Finally, the pipeline from biological data to a mathematical model is demonstrated by addressing questions like: "How comes that hairs are evenly spread over the arm?", "What are the first cell fate decisions?", and "What do have cell clusters in common with penguins?”.

In their early days, banks made a business based on two pillars: Pay interest in exchange for a money deposit on the one side and provide loans in exchange for interest payments on the other. Then came investment banks, computers, brokers, and lots of regulations, which increased complexity of the banking business massively.

Consequently, the financial world is full of challenging aspects like predicting investor behavior, estimating default probabilities, or the valuation of complex financial products. As in natural sciences, this makes it crucial to find or develop the fitting model to a problem, calibrate carefully, and choose parameters wisely.

Such important tasks are very suitable for d-fine consultants, because their academic career as natural scientists has given them the required analytical background. They drill down on problems to understand all relevant components. They abstract in order to find a proper solution. They know how to deal with data. They can code. And although they do not have an economic education, they often find a solution where others have already given up.

Marc Henniges received his PhD in physics from Goethe University Frankfurt in 2012 and has since been working as a financial consultant at d-fine. He will explain what it is like to work in a team of former scientists: What is different to a career in science, and what is similar? How can scientific skills be applied in daily work? And what else could be waiting for you on the career path as financial consultant?

d-fine is a quantitative, IT and risk management consultancy with more than 800 consultants across Europe. They offer quantitative, technical, and technological excellence to their clients, e.g. banks, insurance companies and industrial corporations. Working with d-fine can typically involve dealing with innovative financial products, developing efficient processes, building high performance IT systems or using sophisticated mathematical models to quantify multiple sources of risks.