## Talks in Computational Science

**Speaker : Tamar Schlick, New York University**

**Title : Adventures in Computational Biology: Modeling and Applications**

**Abstract : With significant software and hardware advances, molecular dynamics (MD) simulations have become important for studying the motions of complex biological systems. For various DNA polymerases in the X family, classical as well as classical/quantum-mechanical simulations have uncovered conformational pathways to relate enzyme architecture to fidelity behavior. Going beyond MD, however, is necessary to capturing large-scale conformational changes and chemical pathways. Such methods include transition path sampling and coarse grained modeling approaches. For DNA polymerase beta, transition path sampling and hybrid classical/quantum approaches help relate free energy pathways to biological function. Studies of pol lambda and pol X elucidate the distinct pathways of these polymerases from each other and from pol beta. Applications to chromatin folding require a drastic reduction of the number of degrees of freedom by a coarse-grained approach. Using such a model of oligonucleosome chains in combination with tailored sampling protocols, we elucidate the energetics of oligonucleosome folding/unfolding and the role of each histone tail, linker histones, linker DNA length, and divalent ions in regulating chromatin structure. The resulting compact topologies reconcile features of the zigzag model with straight linker DNAs with the solenoid model with bent linker DNAs foroptimal fiber organization.**

————————————————————————————–

**Speaker : Francois Gygi – University of California, Davis**

**Title : New Scalable Algorithms and Implementations for Large-Scale Molecular Simulations**

**Abstract : First-Principles Molecular Dynamics (FPMD) is a widely used simulation method with applications to Materials Science, Physics and Chemistry. Its popularity is in good part due to its unique ability to describe structural, dynamical and electronic properties in a seamless and consistent way. As large-scale parallel computers become more commonly available, the efficiency of FPMD implementations becomes critical in order to extend the scale and accuracy of first-principles simulations. This however requires adapting and redesigning numerical algorithms in order to achieve good scalability and to make efficient use of multi-core processors and in some cases graphics processing units (GPUs). We present recent progress in the development of parallel numerical linear algebra algorithms used in FPMD for operation on large parallel platforms. Applications to calculations of the electronic structure of nanoparticles are used to illustrate the challenges encountered when running large-scale simulations.**

————————————————————————————–

**Speaker : Romaine Teyssier, University of Zurich**

**Title: High Performance Computing and Galaxy Formation**

**Abstract: The current theory of galaxy formation has been quite successful in explaining the distribution and the basic properties of the billions of galaxies that populate our universe, including our own, the Milky Way. Still, some lingering problems remain: what is the nature of dark energy ? what is dark matter ? why do we see spiral and elliptical galaxies ? These mysteries still elude us, and solutions are probably hidden behind the cosmic complexity of our universe.High performance computing is a key tool in our attempt to understand this complexity. In the last 20 years, computational cosmology has expanded considerably, with larger and larger simulations, including more and more realistic physical models. Using advanced algorithms based on a robust mathematical methodology, new massively parallel codes have been developed, among these the community code RAMSES. Thanks to these technical developments and to the ever increasing size of supercomputers, more robust and accurate predictions are made possible, and sometimes a spectacular computational discovery ! With the advent of Petascale computers, and the constant upgrade of our codes, high performance computing will play a central role in the quest for dark energy and in our understanding of the Milky Way.**

————————————————————————————–

**Speaker : Nikolaus Adams – TU Munich**

**Title : Flows with Interfaces – a Case Study for High-Performance Modeling and Simulation in Engineering**

**Abstract: Predictive simulations in engineering currently face a change in paradigm. Until recently, simulation has been employed mainly for the detail optimization of an existing product while new products were explored by ̶ often extremely expensive – experimental campaigns. The complexity of current technological needs, however, requires the use of explorative computer simulation as tool, already at early stages of product design. As for complex technological problems even the computational modeling itself becomes explorative, the triad “engineering, modeling and computing” no longer can be seen as a sequential cyclic process but now is strongly coupled. The new demands best can be described as “high-performance modeling and computation” (HPMC), where it becomes evident that the need for accuracy and high computational performance strongly affects modeling and vice versa.**

As case study we will consider the modeling of interfacial transport processes in flows. Single-phase flows already can exhibit a large range of strongly coupled scales whose efficient computational modeling poses a challenge by itself. Immersed interfaces add small-scale transport phenomena and can dominate the overall flow evolution. We will show how the demand for high-performance modeling can be addressed from different aspects and how such high-performance models can be embedded in simulation environments for complex flows. Two examples from current research will illustrate the use of HPMC by industry for exploration. The talk will conclude by outlining a vision for meeting future needs of HPMC.

————————————————————————————-

**Speaker : Eric Mjolsness – University of California, Irvine**

**Title: Computational and Mathematical Methods for Multiscale Biological Modeling**

**Abstract: One way to develop methods for modeling complex systems is through new computer languages for modeling. We have built a nested series of modeling languages of increasing capability for modeling biological cells and the development of multicellular organisms. We have collaboratively applied these tools to solve real scientific problems in bacterial metabolism, transcriptional regulation, and plant development including phyllotaxis. Starting with basic representations of chemical reactions and algebraic expressions, these languages allow one to generate deterministic or stochastic models that can be executed, evolved and analyzed symbolically within a commercial mathematical problem-solving environment; also they can be graphically edited, executed and evolved via the web. However, substantial increases in expressivity are required to model the spatially extended and variable-topology systems that arise naturally in development and in multiscale modeling. Indeed, a new subfield of “ computational morphodynamics” – local dynamics of form in biology and technology – demands such capabilities on a grand scale. Fortunately the reaction-like biological process modeling framework can be extended to include quantitative graph grammars, mechanical structures, spatial field dynamics, and even the evolutionary dynamics of such models. Mathematical methods from physics, topology, machine learning, model reduction, and operator algebras can be used to simulate and analyze such models. The result will be a new paradigm for complex multiscale modeling in scientific computing.**