Motivation: Computational approaches to protein function prediction infer protein function by finding proteins with similar sequence, structure, surface clefts, chemical properties, amino acid motifs, interaction partners or phylogenetic profiles. We present a new approach that combines sequential, structural and chemical information into one graph model of proteins. We predict functional class membership of enzymes and non-enzymes using graph kernels and support vector machine classification on these protein graphs.

Results: Our graph model, derivable from protein sequence and structure only, is competitive with vector models that require additional protein information, such as the size of surface pockets. If we include this extra information into our graph model, our classifier yields significantly higher accuracy levels than the vector models. Hyperkernels allow us to select and to optimally combine the most relevant node attributes in our protein graphs. We have laid the foundation for a protein function prediction system that integrates protein information from various sources efficiently and effectively.

Availability: More information available via www.dbs.ifi.lmu.de/Mitarbeiter/borgwardt.html.

Contact: borgwardt@dbs.ifi.lmu.de

/pdf/protein-function-prediction-via-graph-kernels-4pdghwu2q9.pdf

Protein function prediction via graph kernels

Activation of Pheromone-Sensitive Neurons Is Mediated by Conformational Activation of Pheromone-Binding Protein

Quantitative interpretation and prediction of Hofmeister ion effects on protein processes, including folding and crystallization, have been elusive goals of a century of research. Here, a quantitative thermodynamic analysis, developed to treat noncoulombic interactions of solutes with biopolymer surface and recently extended to analyze the effects of Hofmeister salts on the surface tension of water, is applied to literature solubility data for small hydrocarbons and model peptides. This analysis allows us to obtain a minimum estimate of the hydration b1 (H2O A−2), of hydrocarbon surface and partition coefficients Kp, characterizing the distribution of salts and salt ions between this hydration water and bulk water. Assuming that Na+ and SO42− ions of Na2SO4 (the salt giving the largest reduction in hydrocarbon solubility as well as the largest increase in surface tension) are fully excluded from the hydration water at hydrocarbon surface, we obtain the same b1 as for air-water surface (∼0.18 H2O A−2). Ran...

/pdf/thermodynamic-origin-of-hofmeister-ion-effects-4uss22bo9l.pdf

Thermodynamic Origin of Hofmeister Ion Effects

http://compbio.biosci.uq.edu.au/mediawiki/upload/3/37/Buchete_JMB2005.pdf

Molecular dynamics simulations of Alzheimer's beta-amyloid protofilaments.

Herpesviruses, which cause many incurable diseases, infect cells by fusing viral and cellular membranes. Whereas most other enveloped viruses use a single viral catalyst called a fusogen, herpesviruses, inexplicably, require two conserved fusion-machinery components, gB and the heterodimer gH-gL, plus other nonconserved components. gB is a class III viral fusogen, but unlike other members of its class, it does not function alone. We determined the crystal structure of the gH ectodomain bound to gL from herpes simplex virus 2. gH-gL is an unusually tight complex with a unique architecture that, unexpectedly, does not resemble any known viral fusogen. Instead, we propose that gH-gL activates gB for fusion, possibly through direct binding. Formation of a gB-gH-gL complex is critical for fusion and is inhibited by a neutralizing antibody, making the gB-gH-gL interface a promising antiviral target.

/pdf/crystal-structure-of-the-conserved-herpesvirus-fusion-w3kzvkzcjf.pdf

Crystal structure of the conserved herpesvirus fusion regulator complex gH-gL

New computer programs, SurfRace and FastSurf, perform fast calculations of the solvent accessible and molecular (solvent excluded) surface areas of macromolecules. Program SurfRace also calculates the areas of cavities inaccessible from the outside. We introduce the definition of average curvature of molecular surface and calculate average molecular surface curvatures for each atom in a structure. All surface area and curvature calculations are analytic and therefore yield exact values of these quantities. High calculation speed of this software is achieved primarily by avoiding computationally expensive mathematical procedures wherever possible and by efficient handling of surface data structures. The programs are written initially in the language C for PCs running Windows 2000/98/NT, but their code is portable to other platforms with only minor changes in input-output procedures. The algorithm is robust and does not ignore either multiplicity or degeneracy of atomic overlaps. Fast, memory-efficient and robust execution make this software attractive for applications both in computationally expensive energy minimization algorithms, such as docking or molecular dynamics simulations, and in stand-alone surface area and curvature calculations.

Yuri V. Sergeev

Papers

Novel computer program for fast exact calculation of accessible and molecular surface areas and average surface curvature.