Postal address:  
Institute of Environmental Physics, INF 229
University of Heidelberg
D-69120 Heidelberg
Visit: Im Neuenheimer Feld 229
Room Number: 202
Voice: (+49) 6221 54-6379
Skype: Yunus.Sevinchan

Research Interests

Evolution can be seen as the universal generator of complexity in our environment. Systems that behave evolutionary, in the Darwinian sense, span vastly different time and length scales [Szathmáry, 2015], ranging from pre-biotic evolution in chemical reaction networks [Vasas et al., 2012; Nghe et al., 2015] and the origin of eukaryotes [Koonin, 2014; Martin, 2017] to eusocial animal societies [Wilson & Wilson, 2007], and technological [Solé et al., 2013] and cultural evolution [Brewer et al., 2017]. These systems exhibit emergent behaviour like speciation, cooperation, group formation, and hierarchical organisation [Bourke, 2009]. With evolution playing a role in so many domains, increasing the understanding of these processes remains a very intriguing field of research.

For my doctoral studies I am focussing on questions of group formation in complex evolutionary interaction networks. To that end, I am formulating a model of evolving ecological networks and am studying its behaviour through numerical simulations [Brännström, 2012; Landi et al., 2018]. The network includes trophic as well as mutualistic interactions and is embedded in a non-trivial environment. It takes into account the resource flow through the network, the utilisation by its entities, their non-trophic interactions as well as effects mediated or forced by the environment, while the network structure changes adaptively.

Recently, Kotil and Vetsigian [2018; Kelsic et al., 2015] have shown how evolutionary stable communities can form among microbial species, and Grilli et al. [2017; Bairey et al., 2016] demonstrated the stabilising effect of higher-order interactions in ecological networks. One important goal for my doctoral studies is to elevate these approaches to models of spatially distributed and evolving ecological networks and evaluate whether their conclusions hold in this more complex setting. The resulting model offers a framework to investigate not only group formation, but also isolate eco-evolutionary dynamics and study its resilience against perturbations or long-term environmental change. Furthermore, it allows to explore abstractions and the minimal necessities for a certain phenomenology to take place.

Overall, I have a keen interest in connecting theoretical approaches and model building to observations in research on ecology, human behaviour, and their impact on each other. These observations should be used for re-evaluating the premises of a model, and lead to a better understanding of the underlying mechanisms of evolutionary systems. Ultimately, I hope that research in this field will also lead to a better understanding of the dynamics of System Earth and the role of humankind in it.


Former Project (MSc)

In my master thesis project I used computational modelling to investigate how abstracted populations harvest energy from the resources in their environment. To that end, I implemented an evolutionary interaction network which models complex resource and population dynamics and allows the populations to gain innovations that allow them to tap new resources.

Example of possible network entities. Energy source (yellow); non-renewable and renewable resources (grey); populations (green) tapping those resources; energy sink (red).

Once a population gets to a certain size, it splits into two and the offspring population can pick up mutations both in its internal properties and its innovations, i.e. its abilities to tap certain resources. The below figure shows an example of the resulting evolution of one single population parameter, the size beyond which a population splits:

 Exemplary simulation run showing the evolution of the split size parameter of the populations. The initial value of the starting population is 1000. The lines' extension corresponds to the lifetime of the population; additionally, the colour denotes the time of splitting from the parent population. 

The model was testing ground for the feasibility of modelling an interaction network with both complex propagation dynamics (from node to node) as well as complex internal dynamics (within individual nodes) and evolutionary mechanisms. It is further expanded within my doctoral studies.