Research interests

My research is all about cellular information processing and ‘decision’ making. That is, I’m interested in questions such as:

“How do cells receive and transmit information?”

“How do they process this information to make decisions?”

“How do they implement these decisions into action?”

“How does disease occur if these processes go awry and how can we intervene therapeutically?”

To find answers to those questions, I apply experimental techniques from biochemistry and cell biology. Due to the complexity and non-linearity of biological systems, I also approach them from a theoretical perspective and use mathematical models to study the behavior of signal transduction networks. By integrating data from experiments into mathematical models, it is possible to study how complex functions and behavior depend on individual interactions and system components, thereby generating insights that are difficult to obtain by experiments or theory alone. Often, the simulation of such models leads to unexpected predictions that encourage new experiments and facilitate surprising discoveries.


Small G-Protein signaling networks

Small G-Proteins from the Ras-family (Ras, Rho, Rab, Ran) are signalling proteins which can bind guanosine nucleotides and are in an active conformation when bound to GTP, but inactive when bound to GDP. When active, small G-Proteins control a large variety of cellular processes such as gene expression, cytoskeletal remodeling, cell migration and intracellular transport. To ensure faithful regulation of these processes, small GTPase activity itself is strictly controlled both spatially and temporally by GTPase regulatory proteins such GEFs (activating their cognate G-Protein), GAPs (inactiving the G-protein) and GDIs (controlling their localization). Disruption of small GTPase signaling networks is often associated with diseases such as cancer.

My previous research involved the structural validation of novel guanosine nucleotide analoga and developing a novel protocol for the identification of unknown GEFs. Currently, I’m studying the giant muscle protein obscurin with a particular emphasis on its RhoGEF domains and their downstream effects in striated muscle cells within a BHF funded PhD project in the Randall Centre of Cell and Molecular Biophysics at King’s College London.

Phospholamban and The Role of Homo-Oligomers in BIOCHEMICAL SIGNALING NETWORKS

Approximately 30-50% of all proteins in vertebrates are estimated to be able to form homo-oligomers, i.e. supramolecular protein complexes consisting of multiple identical subunits. Important examples of homo-oligomeric signalling proteins include cell surface receptors (growth factor receptors, GPCRs, …), the “guardian of the genome” p53, transcription factors, small GTPases of the Ras family and many more. Yet, the specific advantage of homo-oligomerisation remains poorly understood in many cases.

Based on mathematical models of the oligomerisation reaction, I demonstrated showed that homo-oligomerisation could allow for dynamic signal encoding and homeostatic regulation of monomer concentrations without requiring feedback. By providing pseudo-multisites for post-translational modifications, homo-oligomers could also facilitate ultrasensitive and bistable output behaviour. This indicates that homo-oligomers could play a far greater variety of regulatory functions in signal transduction than previously appreciated.

The micro-peptide phospholamban is a particularly fascinating homo-oligomer. Phospholamban can be phosphorylated by protein kinase A which increases calcium reuptake by the SERCA pump in cardiomyocytes, making phospholamban an important component of β-adrenergic signaling cascade and mediator of the “fight or flight” response. Although phospholamban is known to exist in a constitutive monomer-pentamer equilibrium, the physiological role of pentamers is long-standing open question which I’m currently trying to address from a dynamical systems perspective by using mathematical and experimental approaches.

Bayesian Networks and Cellular automata

Much of my theoretical research involves the use of ordinary differential equations (ODEs) to describe a system mathematically. ODEs are a very useful (and by far the most common) approach to study dynamical systems but they lack spatial information and large models can be difficult to simulate, to analyze and to calibrate. However, many other mathematical modelling approaches exist. Bayesian networks, for example, are a class of probabilistic graphical models particularly well suited for reasoning under uncertainty and for which powerful machine learning algorithms exist. My collaborator Alexander Gebharter and I used causal Bayesian networks to study potential disease progression mechanism in chronic myeloid leukemia and found that an increase in oncogenic Bcr-Abl levels alone is insufficient to explain disease progression and that positive feedback loops and secondary alterations such as additional mutations are likely required before the disease progresses.

Currently we are developing hybrid methods by combining Bayesian networks with agent based approaches such as cellular automata and explore potential applications.