Biomolecular force fields I

• David van der Spoel (Uppsala University)

Room: 132:028 Classical non-polarizable force fields. Computing physical properties that can be compared to experiment. Introducing quantum corrections.

Advanced Monte Carlo Methods

• Ulrich Hansmann (Department of Chemistry and Biochemistry, University of Oklahoma)

Room: 132:028 Ensemble averages vs. time averages, improved updates; reweighting techniques; generalized ensembles, parallel tempering techniques, optimization techniques, recent applications, limitations and future directions.

DNA and DNA-protein interaction

• Richard Lavery (BMSSI, CNRS / Univ. Lyon I)

Room: 132:028 All-atom molecular dynamics applied to DNA and DNA-protein complexes.

Biomolecular force fields II

• David van der Spoel (Uppsala University)

Room: 132:028 State-of-the-art in protein force fields and simulation. Comparison of long simulations with different force fields. Large and long simulations of biomolecules.

Coarse-grained models for proteins I

• Adam Liwo (University of Gdansk)

Room: 132:028 1. Purpose and overview of history and applications of coarse-grained force fields for proteins. 2. Representation of polypeptide chains in coarse-grained force fields. 3. Connection between statistical mechanics and coarse-grained force field 4. Types of potentials (physics-based, statistical, structure-based, engineered, elastic network). 5. Coarse-grained force fields from potentials of mean force: emergence and role of multibody terms. 6. Equations of motion with coarse-grained force field and algorithms for their solving. 7. Generalized-ensemble sampling with coarse-grained force field. 8. Examples (UNRES, CABS, MARTINI).

Long-Timescale Molecular Dynamics Simulations of Protein Folding and Dynamics

• Kresten Lindorff-Larsen (Copenhagen University)

Room: 132:028 Simulations of protein folding. Testing force fields with folding simulations. Comparison of simulations and experiments. Long-timescale motions in folded proteins. Conformational clustering. Dynamics in unfolded proteins.

Parameterisation aspects of atomistic and coarse-grained models of biomolecules

• Alessandra Villa (Karolinska Institutet)

Room: 132:028 Parametrization strategy in biomolecular atomistic force field. Design strategies to build a coarse grained model (mapping scheme, potentials, solvent description). Approach used to parametrize CG potentials (with attention to non-bonded interactions). Backmapping. Transferability problems. Example from a fragment-based coarse grained model for peptide. Brief introduction of possible approach to access multi-scale modeling.

Nucleation in peptide systems

• Stefan Auer (Centre for Molecular Nanoscience, University of Leeds)

Room: 132:028 Calculation of the peptide phase diagram (J. Chem. Phys. 135, 175103 (2011)). Introduce the tube model and the Monte Carlo simulations related to determine the solubility diagram. Nucleation. Numerical work. Application of atomistic nucleation theory to describe amyloid nucleation.

Coarse-grained models for proteins II

• Adam Liwo (University of Gdansk)

Room: 132:028 continued

The internal dynamics of proteins I

• Cristian Micheletti (SISSA)

Room: 132:028 Coarse-grained modelling of proteins' internal dynamics. Generalities about functionally-oriented large-scale structural fluctuations in proteins. Modeling proteins' internal dynamics using elastic network models (ENMs): Stochastic diffusion of a free particle. Stochastic motion of an harmonic oscillator.

The internal dynamics of proteins II

• Cristian Micheletti (SISSA)

Room: 132:028 Stochastic motion of a set of coupled harmonic oscillators. Applications to proteins' internal dynamics. Principal component analysis of MD trajectories.

Hydrodynamic and mesoscopic simulations I

• Jens Harting (TU Eindhoven)

Room: 132:028 Hybrid methods including molecular dynamics for solved particles/molecules and (mesoscopic) methods for hydrodynamic solvent interactions. Examples for the latter are Dissipative Particle Dynamics, Multi Particle Collision Dynamics, Lattice Boltzmann. The methods will be introduced and advanced applications including for example multiphase solvents or complex particle interactions will be explained.

The internal dynamics of proteins III

• Cristian Micheletti (SISSA)

Room: 132:028 Comparing the internal dynamics of proteins with different fold: dynamics-based protein alignment.

Hydrodynamic and mesoscopic simulations II

• Jens Harting (TU Eindhoven)

Room: 132:028 continued

Coarse-grained simulations of DNA in confined geometries

• Cristian Micheletti (International school for Advanced Studies (SISSA), Trieste, Italy)

Room: 132:028 The packing of DNA inside bacteriophages arguably yields the simplest example of genome organisation in living organisms [1, 2]. An indirect indication of how DNA is packaged is provided by the detected spectrum of knots formed by DNA that is circularised inside the P4 viral capsid [3, 4]. The experimental results on the knot spectrum of the P4 DNA are compared to results of coarse-grained simulation of DNA knotting in confined volumes. We start by considering a standard coarse-grained model for DNA which is known to be capable of reproducing the salient physical aspects of free (unconstrained) DNA [5]. Specificallty the model accounts for DNA bending rigidity and excluded volume interactions. By subjecting the model DNA molecules to spatial confinement it is found that confinement favours chiral knots over achiral ones, as found in the P4 experiments. However, no significant bias of torus over twist knots is found, contrary to what found in P4 experiments [6, 7]. A good agreement with experiment is found, instead, upon introducing an additional interaction potential that accounts for tendency of contacting DNA portions to order as in cholesteric liquid crystals. Accounting for this local potential allows us to reproduce the main experimental data on DNA organisation in phages, including the cryo-EM observations and detailed features of the spectrum of DNAknots formed inside viral capsids. The DNA knots we observe are strongly delocalized and, intriguingly, this is shown not to interfere with genome ejection out of the phage [8]. [1] Earnshaw WC, Harrison SC (1977) DNA arrangement in isometric phage heads. Nature 268:598-602. [2] Gelbart WM, Knobler CM (2009) Virology. pressurized viruses. Science 323:1682-1683. [3] Arsuaga J, Vazquez M, Trigueros S, Sumners D, Roca J (2002) Proc Natl Acad Sci U S A 99:5373-5377. [4] Arsuaga, J et al. (2005) Proc Natl Acad Sci U S A 102:9165-9169. [5] Rybenkov VV, Cozzarelli NR, Vologodskii AV (1993) Proc Natl Acad Sci U S A 90:5307-5311. [6] Micheletti C, Marenduzzo D, Orlandini E, Sumners DW (2006) J Chem Phys 124:64903-64903. [7] Micheletti C, Marenduzzo D, Orlandini E, Sumners DW (2008) Biophys J 95:3591-3599. [8] Marenduzzo D, Orlandini E, Stasiak A, Sumners DW, Tubiana L, Micheletti C (2009) Proc Natl Acad Sci U S A 106:22269-22274.

PaLaCe: a coarse-grain model for studying the mechanical properties of proteins

• Marco Pasi (BMSSI, Lyon)

Room: 132:028 We present a new coarse-grain protein model PaLaCe (Pasi-Lavery-Ceres) that has been developed to allow rapid studies of protein mechanics and to build up a deeper understanding of the links between mechanics and function. PaLaCe uses an intermediate level protein representation with two or three pseudoatoms per amino acid. Adding explicit peptide groups and backbone hydrogen bonding allows changes in secondary structure to be treated. The PaLaCe force field is composed of physics-based bonded and non-bonded interactions, combined with an implicit solvent term. The force field was parameterized using Boltzmann inversion of the probability distributions derived from a large database of well-resolved protein structures, and then optimized by fitting simulated and experimental distributions using an iterative refinement technique. PaLaCe has been implemented in the MMTK simulation package and can be used for energy minimization, normal mode calculations and molecular or stochastic dynamics. We illustrate its performance by simulating the forced unfolding of a titin immunoglobin domain.

Challenge to design new computational methodology: how to unlock cellular phenomena over extensive scales in time and space

• Ilpo Vattulainen (Tampere University of Technology, Aalto University School of Science and Technology)

Room: 132:028 Instead of making incremental steps in science, one should aim for breakthroughs that really make a difference. That is, while there are numerous interesting topics to explore, how many of them are really important. One logical way to approach this question is to think of biologically relevant phenomena that due to limited computing power cannot be unresolved today, but whose detailed considerations will be within reach, say, 2-5 years from now. Considering this, what are the main limitations in current simulation models and approaches that we should fix in order to make a difference, to clarify some of the related grand questions. In this contribution, I would like to use the opportunity to create some discussion in this spirit, examples of current limitations including hydrodynamics, non-equilibrium, and crowding.

Entropic tension in crowded membranes

• Martin Lindén (Stockholm University)

Room: 132:028 Unlike their model membrane counterparts, biological membranes are richly decorated with a heterogeneous assembly of membrane proteins. These proteins are so tightly packed that their excluded area interactions can alter the free energy landscape controlling the conformational transitions suffered by such proteins. For membrane channels, this effect can alter the critical membrane tension at which they undergo a transition from a closed to an open state, and therefore influence protein function in vivo. Despite their obvious importance, crowding phenomena in membranes are much less well studied than in the cytoplasm. Using statistical mechanics results for hard disk liquids, we show that crowding induces an entropic tension in the membrane, which influences transitions that alter the projected area and circumference of a membrane protein. As a specific case study in this effect, we consider the impact of crowding on the gating properties of bacterial mechanosensitive membrane channels, which are thought to confer osmoprotection when these cells are subjected to osmotic shock. We find that crowding can alter the gating energies by more than 2kT in physiological conditions, a substantial fraction of the total gating energies in some cases. Given the ubiquity of membrane crowding, the nonspecific nature of excluded volume interactions, and the fact that the function of many membrane proteins involve significant conformational changes, this specific case study highlights a general aspect in the function of membrane proteins.

Fullerene interaction with lipid membranes: atomistic and coarse-grained simulation studies

• Luca Monticelli (INSERM, Paris)

Room: 132:028 Biological membranes compartmentalize cells and form the interface between the cell and its environment. Lipid bilayers are fundamental components of cell membranes. Due to their fluidity, it is very difficult to obtain experimentally atomic level structural information on lipid bilayers in their physiologically relevant state. One property that is difficult to explore in experiments is the membrane ability to dissolve different solutes, including transmembrane peptides and synthetic compounds. We investigated the solubility of fullerene in lipid bilayers, and compared it to its solubility in alkanes. Fullerenes and their derivatives have unique properties that make them interesting for a number of technological applications. Moreover, they are biologically active and can enter easily liposomes and different kinds of cells. Despite numerous studies on both synthetic and biological systems, it is yet unclear how these materials interact with lipid bilayers, and their aggregation in membranes is controversial. I will present results on the validation of all-atom models for C60 fullerene, and on the development of a coarse-grained (CG) model compatible with the MARTINI CG force field for lipids and proteins [1-2]. Using both unbiased and non-equilibrium MD techniques, we characterize the thermodynamics and the kinetics of fullerene aggregation in lipid bilayers and in alkanes. We find that, despite the apparent similarity between alkanes and the bilayer interior, membranes are much better solvents for fullerene. Our results are compatible with experiments showing small perturbations of membrane properties upon addition of fullerene. [1] SJ Marrink et al., J Phys Chem B, 111 (2007) 7812 [2] L Monticelli et al., J Chem Theory Comput, 4 (2008) 819

Coarse grained simulations of lipid bilayers

• Friederike Schmid (Mainz University)

Room: 132:028

Monte Carlo studies of protein aggregation

• Anders Irbäck (Computational Biology and Biological Physics, Lund University)

Room: 132:028 The disease-linked amyloid β and α-synuclein proteins are currently subject to intense research. I will discuss ongoing studies, where we use implicit solvent all-atom Monte Carlo methods to explore the conformational ensembles sampled by these proteins. We study the full-length forms with 42 and 140 residues, respectively, and compare our results with existing experimental data. The aim is to identify and characterize conformational mechanisms involved in aggregation, and gain insight into the effects of, for instance, mutations and aggregation-inhibiting small molecules. I will also discuss a study of oligomer growth for a fibril-forming 6-residue fragment of protein tau, based on the same methods.

Modelling in molecular nanoscience

• Stefan Auer (Centre for Molecular Nanoscience, University of Leeds)

Room: 132:028

Parametrizing polarizable force fields based on the induced point dipole model

• Emppu Salonen (Aalto University)

Room: 132:028

Predictive power of computational chemistry - Do you get what you are paying for?

• David van der Spoel (Uppsala University)

Room: 132:028 There is a plethora of different computational chemistry methods available for studying molecules. Within the field of quantum chemistry these are sometimes called "levels of theory", suggesting that these levels correspond to quality. For empirical models no such classification exists, even though researchers in the field usually have an opinion about the relative merits of different methods that may or may not be based on facts, rumors or prejudice. In this lecture I will try to provoke some discussion by showing some as of yet unpublished results combined with results from the literature.

What is Temperature? Microcanonical Approach to the Statistical Mechanics of Molecular Systems

• Michael Bachmann (The University of Georgia)

Room: 132:028 Folding and aggregation of molecules, as well as the adsorption of soft organic matter to solid inorganic substrates belong to the most interesting challenges in studies of structure formation and function of complex macromolecules. The substantially grown interest in the understanding of basic physical mechanisms underlying these processes is caused by their impact in a broad field that ranges from the molecular origin of the loss of biological functionality as, for example, in Alzheimer's disease, to the development of nanotechnological applications such as biosensors. Most of these systems are necessarily of finite size, but molecular structure formation exhibits cooperative effects that resemble similar processes in thermodynamic phase transitions. Inspired by the fact that the density of states, and with it the microcanonical entropy, is the natural result of any generalized-ensemble Monte Carlo simulation, we have introduced a method that allows for a systematic and unique identification and Ehrenfest-like classification of structural transitions in small systems by means of microcanonical analysis. This computational approach to phase transitions, which is hardly accessible in theoretical studies, is particularly useful for the analysis of cooperative behavior in folding, aggregation, and adsorption processes of polymers and proteins. In this talk, I am going to discuss background and application of this method.

SDAFlex: Simulating flexible macromolecules with Brownian dynamics

• Michael Martinez (Heidelberg Institute for Theoretical Studies)

Room: 132:028 Studies of macromolecular interactions in solution are important for understanding biological activities such as protein-protein interactions in the regulation of signaling pathways. Brownian dynamics simulations are well adapted to the computation of kinetic rates of association between two or more macromolecules (often proteins). Furthermore they have been successfully used to perform protein-protein docking and to study protein-surface interactions and crowded macromolecular environments. However, Brownian dynamics simulations are often limited by the representation of macromolecules as rigid bodies. We have addressed this limitation by extending the SDA6 (Simulation of Diffusional Association [Gabdouline, Wade, 1997]) software to incorporate flexibility of interacting macromolecules. The software incorporates features from the original SDA6 software as well as the SDAMM (SDAMM [Mereghetti, Gabdouline, Wade, 2010] ) software designed to study crowded macromolecular environments. The new software, SDAFlex, has been written using an object-oriented approach, uses less memory and can be run in parallel on shared-memory architecture hardware. SDAFlex simulates flexibility by switching between predefined macromolecular conformations determined by normal mode analysis, NMR or molecular dynamics. Two schemes for accepting conformational switches are implemented: the first which minimises the total system energy; the second a Monte-Carlo algorithm. SDAFlex enables fast generation of docking poses using multiple-conformations and calculation of kinetic rates of association when flexibility and/or a crowded environments are accounted for.

Approaches to multiscale modeling and design of biological molecules

• Nikolay Dokholyan (University of North Carolina)

Room: 132:028 Some of the emerging goals in biological sciences are to uncover the roles of molecular structure and dynamics in certain cellular processes and the ability to rationally manipulate these processes. Despite recent revolutionary advances in experimental methodologies, we are still limited in our ability to sample and decipher the structural and dynamic aspects of single molecules that are critical for their biological function. Thus, there is a crucial need for new and unorthodox techniques to uncover the fundamentals of molecular structure and interactions. We developed a multiscale approach, utilizing rapid Discrete Molecular Dynamics (DMD) simulations, that allows us to study large- and small-scale conformational dynamics of molecules and molecular complexes. Using this approach we demonstrate the ability to control protein stability as well as manipulate protein allostery with computational protein design.

Atomistic and Coarse Grained Simulations of Viral Capsids

• Christoph Globisch (Max-Planck-Institut für Polymerforschung, Mainz)
• Venkatramanan Krishnamani (Carnegie Mellon University, Pittsburgh)

Room: 132:028 The major protective coat of most viruses is a highly symmetric protein capsid that forms spontaneously from many copies of identical proteins. Structural and mechanical properties of several such capsids, as well as their self-assembly process, have been studied experimentally and theoretically, including modeling efforts by computer simulations on various scales. Atomistic models include specific details of local protein binding but are limited to small time- and length-scales, while coarse grained (CG) models capture the scales to study protein assembly but often lack the specific local interactions. Multiscale models aim at bridging this gap by systematically connecting different levels of resolution. We have started to develop a multiscale simulation approach to study the protein capsid complex of the Cowpea Chlorotic Mottle Virus (CCMV), a plant virus with an icosahedral symmetric (T=3) shell of 180 identical proteins. Here, we link simulations at different levels of resolution by parameterizing CG models using atomistic simulations of monomers. From this CG level, we predict emergent properties of larger aggregates, which are possible intermediates in the assembly process or otherwise relevant for the mechanical stability of the virus shell. Atomistic (united atom) molecular dynamics simulations in aqueous solution were carried out to study the conformations sampled by these aggregates (on the limited timescale that is accessible to these simulations) and to investigate the interactions at the protein interface. On the CG side we have used and refined two types of models, the MARTINI model (3-4 heavy atoms per CG bead, explicit water representation) [1] and a recently developed implicit solvent protein model by Bereau and Deserno [2]. [1] Marrink, S. J., Risselada, H. J., Yefimov, S., Tieleman, D. P., and de Vries, A. H. (2007) The MARTINI force field: coarse grained model for biomolecular simulations, J Phys Chem B 111, 7812-7824. [2] Bereau, T., and Deserno, M. (2009) Generic coarse-grained model for protein folding and aggregation, J Chem Phys 130, 235106.

Intrinsically Disordered Proteins At Work: Coupled Folding-Binding in a Simple Hydrophobic/Polar Model

• Arnab Bhattacherjee (CBBP, Lund University)

Room: 132:028 Recent advances in molecular biology have revealed that many proteins do not fold spontaneously into stable native state, instead exist in a highly dynamic state without a specific structure. These disordered proteins, often termed as Intrinsically Disordered or Natively Unstructured Proteins, play a pivot role in various cellular mechanisms by interacting with different proteins. However, little is known about their binding mechanism in atomistic details. With a simple but atomistically detailed protein model, we explore the free-energy landscape of pairs of interacting sequences and how it is impacted by 1), variations in the binding mechanism; and 2), the addition of disordered flanks to the binding region. In particular, we focus on various small helical systems and study how different modes of associations impact kinetic and thermodynamic aspects of the interaction.

Protein Folding without Homology or Machine Learning Techniques

• Karl Freed (James Franck Institute, University of Chicago)

Room: 132:028 Successful methods for predicting protein structure from the amino acid sequence have relied upon machine learning methods and the homology to sequences of proteins with known structures. These methods fail when homology is low, when templates are unavailable for large inserts and/or end portions (InsEnds), and when the proteins become large or have multiple domains. Thus, fully “ab initio” methods without homology or machine learning are necessary to provide the concepts and tools to attack these major unsolved problems. We develop an ab initio, iterative Monte Carlo simulated annealing method for sequentially assigning secondary structure and for prediction the overall protein structure. This ItFix method provides structures almost comparable with homology modeling when homology is adequate, but ItFix fares well for sequences with low homology and for InsEnds with as many as 45 amino acids and secondary structure. Predictions are also generated for the folding sequence.

Exploring the landscape for protein folding: from function to molecular machines

• Jose Onuchic (Center for Theoretical Biological Physics, University of California at San Diego)

Room: 132:028 Globally the energy landscape of a folding protein resembles a partially rough funnel with reduced energetic frustration. A consequence of minimizing energetic frustration is that the topology of the native fold also plays a major role in the folding mechanism. Some folding motifs are easier to design than others suggesting the possibility that evolution not only selected sequences with sufficiently small energetic frustration but also selected more easily designable native structures. The overall structures of the on-route and off-route (traps) intermediates for the folding of more complex proteins are also strongly influenced by topology. Going beyond folding, the power of reduced models to study the physics of protein assembly, protein binding and recognition, and larger biomolecular machines has also proven impressive. Since energetic frustration is sufficiently small, native structure-based models, which correspond to perfectly unfrustrated energy landscapes, have shown to be a powerful approach to explore larger proteins and protein complexes, not only folding but also function associated to large conformational motions. Therefore a discussion of how global motions control the mechanistic processes in the ribosome and molecular motors will be presented. For example, this conceptual framework is allowing us to envisage the dynamics of molecular motors from the structural perspective and it provides the means to make several quantitative predictions that can be tested by experiments.

Design and Prediction of Protein Self-assembly

• Ingemar André (Center for Molecular Protein Science, Lund University)

Room: 132:028 Many of the largest protein complexes in biology are composed of a single type of subunit that is repeated a large number of times to generate a functional assembly. Such homomeric structures are often assembled spontaneously from individual components through the process of self-assembly. Research in our group is focused on the prediction of the three-dimensional structure of homomeric assemblies and the rational design of novel self-assembling proteins and peptides. Over the last several years we have developed computational methods to model the structure of homomeric assemblies using the powerful constraint of molecular symmetry. In this presentation I will illustrate how these prediction methods, in conjunction with limited experimental constraints, can be used to tackle important problems in structural biology. The second part of the talk will deal with the rational design of self-assembling proteins and peptides. We combine the powerful design template of self-assembly with structural modeling and computational protein to design protein assemblies on an atomic level. The final part of my talk will deal with open questions relating to protein and peptides self-assembly that I am interested in exploring during the workshop. In particular, I am interested in questions relating to the evolution of protein building blocks capable of complex self-assembly, the assembly mechanism of multiprotein complexes and the fine-tuning of intermolecular interactions in protein assemblies.

Cooperativity, Local-Nonlocal Coupling, and Nonnative Interactions in Protein Folding

• Hue Sun Chan (Departments of Biochemistry, of Molecular Genetics, and of Physics, University of Toronto)

Room: 132:028 The Levinthal paradox of protein folding is commonly perceived as a statement about the impossibility of folding by a completely random conformational search. Often missed in such narratives is the fact that the question raised by Levinthal was in response to the experimental discovery of two-state, switch-like cooperative folding in the late 1960s, rather than to the problem of conformational search per se. The implication of this understanding on the notion of a funnel-like energy landscape will be discussed. Comparisons between theory and experiment on cooperative folding indicate a prominent role of desolvation barriers, which contribute to an apparent general organizing principle entailing a coupling between local conformational preferences and nonlocal packing interactions. Investigations into the role of desolvation in protein folding also resolves an apparent inconsistency between experimental observations of enthalpic folding barriers and the theoretical funnel picture of folding. Examples will be given to illustrate how important folding principles have been gleaned from studies using native-centric models and how nonnative interactions may be treated perturbatively in essentially the same conceptual framework.

Computer simulation of protein-protein association

• Volkhard Helms (Saarland University, Center for Bioinformatics)

Room: 132:028 Spaar, A. and Helms, V. (2005) Journal of Chemical Theory and Computation, Vol. 1 (4), p. 723-736. Free Energy Landscape of Protein-Protein Encounter Resulting from Brownian Dynamics Simulations of Barnase: Barstar Spaar, A., Dammer, C., Gabdoulline, R.R., Wade, R.C., and Helms, V. (2006) Biophysical Journal, Vol. 90, p. 1913-1924. Diffusional Encounter of Barnase and Barstar. Ahmad, M., Gu, W., and Helms, V. (2008), Angewandte Chemie International Edition, Vol. 47, p. 7626-7630. Mechanism of Fast Peptide Recognition by SH3 Domains (EN). Ahmad, M., Gu, W., Geyer, T., and Helms, V. (2011) Nature Communications, Vol. 2, Article no. 261. Adhesive water networks facilitate binding of protein interfaces (Abstract).

Structure and dynamics of lipid bilayers from simulations

• Olle Edholm (KTH)

Room: 132:028

Challenges in protein structure prediction

• Arne Elofsson (Dep of Biochemistry and Biophysics, Stockholm University)

Room: 132:028 Proteins are the central machines of cells, and they perform their actions by interacting with each other as well as with other molecules. Large complexes involving tens or even hundreds of proteins make up the central hubs in biological interaction networks. In human cells repeated domains are frequent among these hubs. Today, large-scale efforts in genomics, proteomics, lipidomics and metabolomics are producing complete lists of the molecules in entire cell as well as in different sub-cellular compartments. Further, interactions between molecules can be studied at different levels of detail. In small-scale studies it is possible to obtain detailed information about the interaction of a few molecules, while in large-scale studies less detailed information for a larger set of molecules can be obtained. Only for a small number of the large complexes atomistic details have been possible to obtain and in particular molecular complexes embedded in the membrane have been difficult to study experimentally. A major aim within the field is to reveal detailed structural information about large biological complexes. To obtain this goal a mix of experimental and computational methods needs to be applied. A major source of information is coming from the rapid increase in genomic sequence data. Here, I will discuss how to combine computational and experimental studies to obtain increased understanding of the formation of large molecular complexes particularly in the membrane.

Probabilistic models of protein structure: from theory to applications

• Thomas Hamelryck (University of Copenhagen)

Room: 132:028 The so-called protein folding problem is the loose denominator for an amalgam of closely related, unsolved problems that include protein structure prediction, protein design and the simulation of the protein folding process. We adopt a unique probabilistic approach to modeling bio-molecular structure, based on graphical models and directional statistics [1,2,3,4,5]. Notably, we developed the first probabilistic model of protein structure in full atomic detail [1, 4]. In this talk, I will give an overview of how rigorous probabilistic models of something as complicated as a protein structure can be formulated, focusing on the use of graphical models and directional statistics to model angular degrees of freedom. I will also discuss the reference ratio method [6], a novel statistical method that can be seen as a surprising Bayesian variant of the maximum entropy method. This method also sheds an entirely new light on the in protein structure prediction widely used potentials of mean force, which were up to now poorly understood and justified. Finally, I will describe some applications, including the investigation of protein dynamics and the statistical inference of protein structure from nuclear magnetic resonance (NMR) data [7] and small angle X-ray scattering (SAXS) data [8]. [1] Hamelryck, T., Kent, J., Krogh, A. (2006) Sampling realistic protein conformations using local structural bias. PLoS Comput. Biol., 2(9): e131. [2] W. Boomsma, K.V. Mardia, C.C. Taylor, J. Ferkinghoff-Borg, A. Krogh, and T. Hamelryck. (2008) A generative, probabilistic model of local protein structure. Proc. Natl. Acad. Sci. U S A, 105(26):8932–8937. [3] Frellsen, J., Moltke, I., Thiim, M., Mardia, KV., Ferkinghoff-Borg, J., Hamelryck, T. (2009) A probabilistic model of RNA conformational space. PLoS Computational Biology, 5(6), e1000406. [4] Tim Harder, Wouter Boomsma, Martin Paluszewski, Jes Frellsen, Kristoffer Johansson, and Thomas Hamelryck. (2010) Beyond rotamers: a generative, probabilistic model of side chains in proteins. BMC Bioinformatics, 11(1):306. [5] M. Paluszewski and T. Hamelryck. (2010) Mocapy++ – a toolkit for inference and learning in dynamic bayesian networks. BMC bioinformatics, 11(1):126. [6] Thomas Hamelryck, Mikael Borg, Martin Paluszewski, Jonas Paulsen, Jes Frellsen, Christian Andreetta, Wouter Boomsma, Sandro Bottaro, and Jesper Ferkinghoff-Borg (2010). Potentials of mean force for protein structure prediction vindicated, formalized and generalized. PLoS ONE, 5(11):e13714. [7] Simon Olsson, Wouter Boomsma, Jes Frellsen, Sandro Bottaro, Tim Harder, Jesper Ferkinghoff-Borg, and Thomas Hamelryck. (2011) Generative probabilistic models extend the scope of inferential structure determination. J. Magn. Reson., 213(1):182-6. [8] Stovgaard, K., Andreetta, C., Ferkinghoff-Borg, J., Hamelryck, T. (2010) Calculation of accurate small angle X-ray scattering curves from coarse-grained protein models. BMC Bioinformatics, 11:429.

Predicting protein structure by solving the inverse Potts problem: a pseudo-likelihood approach

• Erik Aurell (KTH)

Room: 132:028 Inverse statistical mechanics means to determine model parameters (couplings, external fields etc) from observations (one- and two-point correlations, or other data). In the course of an ongoing investigation into methods to do the "inverse Ising" problem we recently found that a pseudo-likelihood method works better than many other alternatives; it is particularly good in the parameter range of strong interactions and few samples [1]. We have recently tried to extend this method to determine amino acid contacts from protein sequences in the same protein family, following the approach of Morcos et al [2]. The relevant model is then a Potts model with 20 or 21 states. We find that here also the pseudo-likelihood provides somewhat better reconstruction of known protein structures [3]. This is joint work with Magnus Ekeberg and Martin Weigt. [1] Erik Aurell, Magnus Ekeberg "Inverse Ising inference using all the data", Physical Review Letters (2012, in press) [arXiv:1107.3536] [2] Faruck Morcos et al, "Direct-coupling analysis of residue coevolution captures native contacts across many protein families", PNAS November 21, 2011 [3] Magnus Ekeberg, Erik Aurell, Martin Weigt (2012, in preparation)

Bridging the gap: Linking molecular simulations and systemic descriptions of a chromatophore vesicle

• Volkhard Helms (Saarland University, Center for Bioinformatics)

Room: 132:028 Geyer, T. and Helms, V. (2006) Biophysical Journal, Vol. 91, p. 927-937. Reconstruction of a Kinetic Model of the Chromatophore Vesicles from Rhodobacter Sphaeroides. Geyer, T. and Helms, V. (2006) Biophysical Journal, Vol. 91, p. 921-926. A Spatial Model of the Chromatophore Vesicles of Rhodobacter Sphaeroides and the Position of the Cytochrome bc1 Complex. Geyer, T., Lauck, F., and Helms, V. (2007) Journal of Biotechnology, Vol. 129, p. 212-228. Molecular Stochastic Simulations of Chromatophore Vesicles from Rhodobacter Sphaeroides. Geyer, T., Mol, X., Blaß, S., and Helms, V. (2010) PLoS ONE 5(11): e14070. Bridging the gap: Linking molecular simulations and systemic descriptions of cellular compartments.

Applications of multiscale modeling and design of biological molecules

• Nikolay Dokholyan (University of North Carolina)

Room: 132:028 Some of the emerging goals in biological sciences are to uncover the roles of molecular structure and dynamics in certain cellular processes and the ability to rationally manipulate these processes. Despite recent revolutionary advances in experimental methodologies, we are still limited in our ability to sample and decipher the structural and dynamic aspects of single molecules that are critical for their biological function. Thus, there is a crucial need for new and unorthodox techniques to uncover the fundamentals of molecular structure and interactions. We developed a multiscale approach, utilizing rapid Discrete Molecular Dynamics (DMD) simulations, that allows us to study large- and small-scale conformational dynamics of molecules and molecular complexes. Using this approach we demonstrate the ability to control protein stability as well as manipulate protein allostery with computational protein design.

How May a Protein Unknot a Knotted DNA? Statistical Physics of Local Inference of Global Topology by Topoisomerases

• Hue Sun Chan (Departments of Biochemistry, of Molecular Genetics, and of Physics, University of Toronto)

Room: 132:028 Closed DNA circles can be unknotted, knotted or linked (catenated). Such topological entanglements of DNA molecules have important impact on biological processes. Topoisomerases are a ubiquitous class of enzymes that pass one DNA segment through another, serving critical biological functions in cellular replication and maintenance of genome stability. Experimentally, type-2 topoisomerases (topo II) can reduce knot population by as much as 90 times and catenane population by ~ 16 times. These observations raise a fundamental question of physical principle: How does a relatively small enzyme discern the global topology of a much larger DNA molecule that it acts upon? Because it seems that topo II can work magic, it has even been likened to Maxwell's demon. This talk addresses the statistical mechanical basis of topo II actions. Using coarse-grained lattice and continuum wormlike chain models, we have elucidated the mathematical basis of the hypothesis that topo II recognize and act at specific DNA juxtapositions. We found that selective segment passage at hooked geometries can reduce knot populations as dramatically as seen in experiments. Selective segment passage also provided theoretical underpinning for an intriguing empirical scaling relation between unknotting and decatenating potentials. Such selective segment passage also accounts for supercoil simplification (narrowing linking number distribution) by topo II. The consistent agreement between theory and experiment argues for topo II actions at hooked or twisted-hooked DNA juxtapositions. Our investigation also highlights a general connection between local geometry and global topology in polymer configurations.

Automated real-space refinement of crystal and cryoEM structures using a realistic backbone move set

• Karl Freed (James Franck Institute, University of Chicago)

Room: 132:028