Workshop on Computational Tools in Macromolecular Modeling

 October 14th, 2015, Britannia Bldg, Tel Aviv University

Lecture Session (Hall 5):
9:00 - 9:10 Opening Remarks
9:10 - 10:00 Sarel Fleishman (Weizmann Inst.) : Computational design of antibodies and enzymes guided by natural sequences and conformations. Abstract.
10:00 - 10:50 Joel Sussman (Weizmann Inst.) : Proteopedia - a Scientific 'Wiki' Bridging the Rift Between 3D Structure and Function of Biomacromolecules. Abstract.
10:50 - 11:10 Coffee Break (Lobby)
11:10 - 12:00 Distinguished Safra Speaker : Shoshana Wodak (VIB and Toronto U.) : Modeling Protein-protein interactions: then and now. Abstract.
12:00 - 12:50 Haim Wolfson (Tel Aviv U.) : Algorithms for the Integrative Modelling of Large Macromolecular Complexes. Abstract.
12:50 - 13:40 Lunch (Lobby)
Hand-on Tutorials (Computer Lab - IN SHERMAN BLDG - NORTH OF BRITANIA BLDG , limited to 50 participants)
13:40 - 15:30 Nir Kalishman (Hebrew U.) : Analysis of cross-linked and mass-spectrometry (XL-MS) data. Abstract. Download material.
15:30 - 15:45 Coffee Break ( Lobby of Britania bldg )
15:45 - 16:30 Dan Cohen (Tel Aviv U.) : Analysis and Interpretation of cryo-EM maps using the UCSF Chimera software package. Abstract.
16:30 - 17:15 Daniel Zaidman (Tel Aviv U.) : Modeling of Large Multimolecular Complexes using MultiFit. Abstract. Download material.

Lecturer Title Abstract
Sarel Fleishman Computational design of antibodies and enzymes guided by natural sequences and conformations Computational protein design has made substantial progress over recent years generating new enzymes, binders, and inhibitors not seen in nature. Among still unsolved challenges are design of new backbones for function. We have developed new algorithms for design of backbones from modular pieces of natural proteins and applied this strategy to design new antibody binders of insulin. The antibodies, which are distant from any natural germ­line antibody by at least 50 mutations, bind at mid­nanomolar dissociation constants and mutations introduced through experimental affinity maturation appear to rigidify the bound conformation. We also developed methods for designing more stable variants of human proteins that can be expressed in high yields in bacterial systems. We successfully applied this method to 4 challenging enzymes and two binding proteins, in all cases obtaining higher yields and stability than the wild­type protein with no loss in biological activity. In the case of human acetylcholine esterase, the designed variant, encoding 51 substitutions relative to wild­type expresses with over 1,000­fold higher yields than wild­type without sacrificing catalytic efficiency.
Joel L. Sussman Proteopedia - a Scientific 'Wiki' Bridging the Rift Between 3D Structure and Function of Biomacromolecules Students and scientists are now able to access 3D images of biomacromolecules both in journal and on the web. However, rather than just relying on text and 2D images to try to understand the function of biomacromolecular structures, a collaborative website called Proteopedia1,2 is a free resource, which links written information & 3D structural images. This wiki web site, displays protein structures & other biomacromolecules interactively. These images are surrounded by descriptive text containing hyperlinks that change the appearance (e.g., views, representations, colors or labels) of the adjacent 3D structure to reflect the concepts discussed in the text. This makes the complex structural information readily accessible and comprehensible, to non-structural biologists. Using Proteopedia, one can easily create descriptions of biomacromolecules linked to their 3D structure, e.g., see a page on the ribosome structure/function, Pages can be viewed on PCs, MACs & LINUX computers and even on iPads (that do not have JAVA), via the molecular viewer JSmol3, e.g., a page on HIV-1 protease, Content is being added by Proteopedia's >3,00 users, in 60 different countries, in a dozen different languages, including Russian, Chinese, Arabic, & Hebrew: (Hebrew). A number of journals & book publishers are using Proteopedia to complement their printed and web papers using Proteopedia's "Interactive 3D Complements" (I3DCs) - see, e.g., Pages for each of the >111,000 entries in the PDB have been automatically created with 'seed' information, and are both intrinsically useful and 'primed' for expansion by users. Scientists & students are invited to request a Proteopedia user account, at no cost, in order to edit existing pages & to create new ones, see:
Shoshana J. Wodak Modeling Protein-protein interactions: then and now I'll review some of the very early protein-ligand and protein-protein docking methods and their application to the elucidation of the quaternary structure change in hemoglobin, and how they gave rise to a thriving area of research. This will be followed by the description of a recent computational analysis of the mechanism and free energy profile of three-dimensional domain swapping reactions in protein homo-oligomers. In this analysis we propose that a large category of domain swapped systems form by a process that does not involve chain unfolding but proceeds by gradual exchange between intra- and intermolecular interactions between the subunits, which described a minimum energy path for the reaction. The computational analysis models both the conformational and dynamic aspects of the process at the atomic scale.
Haim J. Wolfson Algorithms for the integrative modeling of large multimolecular complexes. Modeling large multi-molecular assemblies at atomic resolution is a key task in elucidating cell function. Since, there is no single experimental method, that can deliver atomic resolution structures of such large molecules, hybrid methods, which integrate data from various experimental modalities, are being developed for this task. We have developed several integrative methods (EMatch, MultiFit, 3D-PUZZLE), which combine atomic resolution models of individual assembly components with an electron microscopy map of the full assembly. 3D-PUZZLE also naturally accommodates available chemical cross link (Xlink) data. The input to our algorithms is an intermediate resolution or low resolution (for MultiFit) electron density map of the full assembly, atomic resolution (2 A0) maps of the individual assembly subunits, and, if available, cross link information between some residues of neighboring subunits (an Xlink can be visualized as a loose ~30A0 string connecting two atoms on the surfaces of neighboring subunits). The output is an atomic resolution map of the whole assembly. We shall focus on the performance of the novel 3D-PUZLLE algorithm, which was highly successful and efficient on all the intermediate resolution EM complexes from the 2010 Cryo-EM Modeling Challenge. Remarkably, a 6.8A0 resolution 20S proteasome map, consisting of 28 (structurally homologues) units was modeled at 1.5A0 RMSD from native in about 10 minutes on a Core i7 laptop. In case of missing (or poorly modeled) individual subunits, the method can return partial solutions, thus, enabling interactive modeling. From a purely geometric viewpoint, the task can be viewed as an assembly of a large multiple piece puzzle, where we have relatively accurate models of the individual subunits, and a rough, low resolution scan of the full puzzle volume.
Nir Kalishman Analysis of cross-linked and mass-spectrometry (XL-MS) data Cross-linking and mass-spectrometry is emerging as a powerful tool in the structural elucidation of large protein complexes. This technique is now readily available to the structural biologist, but it is important to understand the details of the computational analysis. In this session I will briefly explain the methodology and describe the raw data that are measured by the mass-spectrometer. I will then describe what are the computational steps that are required in order to get a list of cross-links out of the data. The session will include a review of available software packages and web-servers for such analysis and a demonstration on data measured in my lab.
Dan Cohen Analysis and Interpretation of cryo-EM maps using the UCSF Chimera software package During the last several years, Cryo-Electron Microscopy has become an important method for protein structure elucidation. Cryo-EM map resolutions have been constantly improving, lately reaching near-atomic ones. Following a short introduction to Cryo-EM, we will focus on the analysis and interpretation of such maps. Several methods for modelling atomic structures based on Cryo-EM maps will be reviewed in detail. The session will include extensive hands-on training using the UCSF Chimera application.
Daniel Zaidman Modeling of Large Multimolecular Complexes using MultiFit MultiFit is a computational method for simultaneously fitting atomic structures of components into their assembly density map at resolutions as low as 25 ?. The component positions and orientations are optimized with respect to a scoring function that includes the quality-of-fit of components in the map, the protrusion of components from the map envelope, as well as the shape complementarity between pairs of components. The scoring function is optimized by an exact inference optimizer DOMINO that efficiently finds the global minimum in a discrete sampling space. In this tutorial you will hear about the algorithm and the way to apply it to your target. You will use Multifit to predict a bound as well as an unbound complex. You will also use the Symmetry mode to predict the atomic structure of the GroEL.