Institute for Advanced Simulation (IAS)

Computational Biomedicine (IAS-5 / INM-9)

Structural Systems Biology

Neurotransmission is a central event for brain function. As a whole, it can be viewed as a communication of information from the pre-synaptic neuron to the post-synaptic one. Signal transduction cascades triggered by chemical neurotransmitters involve a myriad of receptors and enzymes. INM-9/IAS-5, while keeping its main focus on molecular level descriptions of neurotransmission components (such as receptors at the post-synaptic membrane), is making an increasing effort to move towards the systemic level. Key (and mostly unanswered) questions include: when an agonist binds to a neuroreceptor, can we predict its downstream effects, such as epigenetics changes and production of ion current carried out by ion channels? How disease-linked mutations impact on the protein interactome of the synaptic space?

To start addressing these fascinating (and yet very challenging) questions, we have initiated an activity in the integration of a molecular with a systems biology level of description of postsynaptic signaling cascades. Unknown key input parameters of the models here may be provided by simulations. These include hybrid quantum mechanics/molecular mechanics (QM/MM) simulations, coarse-grain/MM calculations, all-atom molecular dynamics(MD)-based kinetics calculations, and Brownian Dynamics. The perspective of combining QM/MM, MD simulations and Machine learning tools, together with systems biology, will allow to connect truly High-Performance Computing (HPC) applications to subcellular investigations, relating molecular-level events to real biological responses.

We are also integrating a variety of computational tools, spanning from network modeling and machine-learning, with our expertise in structural bioinformatics and massive protein-protein docking.

Our new system-level research line will be integrated and automatized in computational platform, called here v-SNAP (Virtual human SyNaptosome And Pharmacology). v-SNAP will augment our existing webserver (GoMoDO, Sandal et al. PLoSOne 2013) to structural predictions of receptors located at the postsynaptic membrane, along with the modeling of the underlying signal pathways. v-SNAP might provide information on the impact of protein variants associated with neurological and neuropsychiatric disorders. Kinetic constants form molecular simulation and/or from the literature will feed the signalling cascade models fired by given neurotransmitters, as well as to evaluate the effect of disease associated variants. v-SNAP will also include an automatic pipeline for the modelling of disease-related proteins variants or pathogenic PPI networks.