DeepSite: protein-binding site predictor using 3D-convolutional neural networks.

MOTIVATION: An important step in structure-based drug design consists in the prediction of druggable binding sites. Several algorithms for detecting binding cavities, those likely to bind to a small drug compound, have been developed over the years by clever exploitation of geometric, chemical and evolutionary features of the protein. RESULTS: Here we present a novel knowledge-based approach that uses state-of-the-art convolutional neural networks, where the algorithm is learned by examples. In total, 7622 proteins from the scPDB database of binding sites have been evaluated using both a distance and a volumetric overlap approach. Our machine-learning based method demonstrates superior performance to two other competitive algorithmic strategies. AVAILABILITY AND IMPLEMENTATION: DeepSite is freely available at www.playmolecule.org. Users can submit either a PDB ID or PDB file for pocket detection to our NVIDIA GPU-equipped servers through a WebGL graphical interface. CONTACT: gianni.defabritiis@upf.edu. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

AceCloud: Molecular Dynamics Simulations in the Cloud.

We present AceCloud, an on-demand service for molecular dynamics simulations. AceCloud is designed to facilitate the secure execution of large ensembles of simulations on an external cloud computing service (currently Amazon Web Services). The AceCloud client, integrated into the ACEMD molecular dynamics package, provides an easy-to-use interface that abstracts all aspects of interaction with the cloud services. This gives the user the experience that all simulations are running on their local machine, minimizing the learning curve typically associated with the transition to using high performance computing services.

Reranking docking poses using molecular simulations and approximate free energy methods.

Fast and accurate identification of active compounds is essential for effective use of virtual screening workflows. Here, we have compared the ligand-ranking efficiency of the linear interaction energy (LIE) method against standard docking approaches. Using a trypsin set of 1549 compounds, we performed 12,250 molecular dynamics simulations. The LIE method proved effective but did not yield results significantly better than those obtained with docking codes. The entire database of simulations is released.

Kinetic characterization of fragment binding in AmpC ß-lactamase by high-throughput molecular simulations.

Small molecules used in fragment-based drug discovery form multiple, promiscuous binding complexes difficult to capture experimentally. Here, we identify such binding poses and their associated energetics and kinetics using molecular dynamics simulations on AmpC ß-lactamase. Only one of the crystallographic binding poses was found to be thermodynamically favorable; however, the ligand shows several binding poses within the pocket. This study demonstrates free-binding molecular simulations in the context of fragment-to-lead development and its potential application in drug design.

High-throughput all-atom molecular dynamics simulations using distributed computing.

Although molecular dynamics simulation methods are useful in the modeling of macromolecular systems, they remain computationally expensive, with production work requiring costly high-performance computing (HPC) resources. We review recent innovations in accelerating molecular dynamics on graphics processing units (GPUs), and we describe GPUGRID, a volunteer computing project that uses the GPU resources of nondedicated desktop and workstation computers. In particular, we demonstrate the capability of simulating thousands of all-atom molecular trajectories generated at an average of 20 ns/day each (for systems of approximately 30 000-80 000 atoms). In conjunction with a potential of mean force (PMF) protocol for computing binding free energies, we demonstrate the use of GPUGRID in the computation of accurate binding affinities of the Src SH2 domain/pYEEI ligand complex by reconstructing the PMF over 373 umbrella sampling windows of 55 ns each (20.5 mus of total data). We obtain a standard free energy of binding of -8.7 +/- 0.4 kcal/mol within 0.7 kcal/mol from experimental results. This infrastructure will provide the basis for a robust system for high-throughput accurate binding affinity prediction.

The impact of accelerator processors for high-throughput molecular modeling and simulation.

The recent introduction of cost-effective accelerator processors (APs), such as the IBM Cell processor and Nvidia's graphics processing units (GPUs), represents an important technological innovation which promises to unleash the full potential of atomistic molecular modeling and simulation for the biotechnology industry. Present APs can deliver over an order of magnitude more floating-point operations per second (flops) than standard processors, broadly equivalent to a decade of Moore's law growth, and significantly reduce the cost of current atom-based molecular simulations. In conjunction with distributed and grid-computing solutions, accelerated molecular simulations may finally be used to extend current in silico protocols by the use of accurate thermodynamic calculations instead of approximate methods and simulate hundreds of protein-ligand complexes with full molecular specificity, a crucial requirement of in silico drug discovery workflows.

Energetics of K+ permeability through Gramicidin A by forward-reverse steered molecular dynamics.

The estimation of ion channel permeability poses a considerable challenge for computer simulations because of the significant free energy barriers involved, but also offers valuable molecular information on the ion permeation process not directly available from experiments. In this article we determine the equilibrium free energy barrier for potassium ion permeability in Gramicidin A in an efficient way by atomistic forward-reverse non-equilibrium steered molecular dynamics simulations, opening the way for its use in more complex biochemical systems. Our results indicate that the tent-shaped energetics of translocation of K+ ions in Gramicidin A is dictated by the different polarization responses to the ion of the external bulk water and the less polar environment of the membrane.

Accuracy of the lattice-Boltzmann method using the Cell processor.

Accelerator processors like the new Cell processor are extending the traditional platforms for scientific computation, allowing orders of magnitude more floating-point operations per second (flops) compared to standard central processing units. However, they currently lack double-precision support and support for some IEEE 754 capabilities. In this work, we develop a lattice-Boltzmann (LB) code to run on the Cell processor and test the accuracy of this lattice method on this platform. We run tests for different flow topologies, boundary conditions, and Reynolds numbers in the range Re=6-350 . In one case, simulation results show a reduced mass and momentum conservation compared to an equivalent double-precision LB implementation. All other cases demonstrate the utility of the Cell processor for fluid dynamics simulations. Benchmarks on two Cell-based platforms are performed, the Sony Playstation3 and the QS20/QS21 IBM blade, obtaining a speed-up factor of 7 and 21, respectively, compared to the original PC version of the code, and a conservative sustained performance of 28 gigaflops per single Cell processor. Our results suggest that choice of IEEE 754 rounding mode is possibly as important as double-precision support for this specific scientific application.

Embedding molecular dynamics within fluctuating hydrodynamics in multiscale simulations of liquids.

We present a hybrid protocol designed to couple the dynamics of a nanoscopic region of liquid described at atomistic level with a fluctuating hydrodynamics description of the surrounding liquid. The hybrid technique is based on the exchange of fluxes and it is shown to respect the conservation laws of fluid mechanics. This fact allows us to solve unsteady flows involving shear and sound waves crossing the interface of both domains. In equilibrium we find perfect agreement with the grand-canonical ensemble at low and moderate densities, while within the nanoscopic volumes considered, mass fluctuation (both in hybrid and full MD simulations) becomes slightly larger than predicted by the thermodynamic limit. Stress fluctuations across the hybrid interface are shown to have a seamless profile. Nonequilibrium scenarios involving shear (startup of Couette flow) and longitudinal flow (sound waves) are also illustrated.

Fluctuating hydrodynamic modeling of fluids at the nanoscale.

A good representation of mesoscopic fluids is required to combine with molecular simulations at larger length and time scales [De Fabritiis, Phys. Rev. Lett. 97, 134501 (2006)]. However, accurate computational models of the hydrodynamics of nanoscale molecular assemblies are lacking, at least in part because of the stochastic character of the underlying fluctuating hydrodynamic equations. Here we derive a finite volume discretization of the compressible isothermal fluctuating hydrodynamic equations over a regular grid in the Eulerian reference system. We apply it to fluids such as argon at arbitrary densities and water under ambient conditions. To that end, molecular dynamics simulations are used to derive the required fluid properties. The equilibrium state of the model is shown to be thermodynamically consistent and correctly reproduces linear hydrodynamics including relaxation of sound and shear modes. We also consider nonequilibrium states involving diffusion and convection in cavities with no-slip boundary conditions.

HTMD: High-Throughput Molecular Dynamics for Molecular Discovery.

Recent advances in molecular simulations have allowed scientists to investigate slower biological processes than ever before. Together with these advances came an explosion of data that has transformed a traditionally computing-bound into a data-bound problem. Here, we present HTMD, a programmable, extensible platform written in Python that aims to solve the data generation and analysis problem as well as increase reproducibility by providing a complete workspace for simulation-based discovery. So far, HTMD includes system building for CHARMM and AMBER force fields, projection methods, clustering, molecular simulation production, adaptive sampling, an Amazon cloud interface, Markov state models, and visualization. As a result, a single, short HTMD script can lead from a PDB structure to useful quantities such as relaxation time scales, equilibrium populations, metastable conformations, and kinetic rates. In this paper, we focus on the adaptive sampling and Markov state modeling features.

[Foreign bodies in the alimentary tract. Surgical treatment]. / I corpi estranei del canale alimentare. Trattamento chirurgico.

Gastro-intestinal foreign bodies are a by no means rare event in surgery and in the USA mortality is about 1500 people per annum. The surgical treatment of foreign bodies in the alimentary tract is reported here. Certain cases of voluntary ingestion in mental patients are reported, comparing personal experience with the data reported in the literature.

[Cancer of the colorectum in the aged. Clinical considerations with reference to our case series of 65 operated cases during the triennial, 1984-1986]. / Il cancro del colon retto nell'anziano. Considerazioni cliniche in riferimento alla nostra casistica di 65 casi operati nel triennio 1984-1986.

Starting from the note that in industrialised countries colorectal tumours are an increasingly serious problem, especially in the elderly, and after some epidemiological remarks, a personal series of 65 consecutive operations on over--70s in a three-year period is considered. Personal statistics are analysed following careful assessment of risk factors and the immediate and long-term surgical results, also examined on the basis of reported data. It is noted, first, that age is never an absolute contraindication to surgery; second that early diagnosis is basic for the achievement of an improved prognosis: proof of this lies in the excessive number of emergency operations for occlusion or perforation. On the other hand, while it is true that extreme radicalism at an advanced stage does not imply any substantial modification to prognosis, it should also be recognised that the shortening of surgical times (after the introduction of mechanical staplers and the improvement in anaesthesiological assistance techniques) offer greater scope for manoeuvre.

On-the-Fly Learning and Sampling of Ligand Binding by High-Throughput Molecular Simulations.

High-throughput molecular dynamics (MD) simulations are a computational method consisting of using multiple short trajectories, instead of few long ones, to cover slow biological time scales. Compared to long trajectories this method offers the possibility to start the simulations in successive batches, building a knowledgeable model of the available data to inform subsequent new simulations iteratively. Here, we demonstrate an automatic, iterative, on-the-fly method for learning and sampling molecular simulations in the context of ligand binding for the case of trypsin-benzamidine binding. The method uses Markov state models to learn a simplified model of the simulations and decide where best to sample from, achieving a converged binding affinity in approximately one microsecond, 1 order of magnitude faster than classical sampling. This method demonstrates for the first time the potential of adaptive sampling schemes in the case of ligand binding.

Molecular modeling and simulation of membrane lipid-mediated effects on GPCRs.

Functioning of G protein-coupled receptors (GPCRs) is tightly linked to the membrane environment, but a molecular level understanding of the modulation of GPCR by membrane lipids is not available. However, specific receptor-lipid interactions as well as unspecific effects mediated by the bulk properties of the membrane (thickness, curvature, etc.) have been proposed to be key regulators of GPCR modulation. In this review, we examine computational efforts made towards modeling and simulation of (i) the complex behavior of membrane lipids, (ii) membrane lipid-GPCR interactions as well as membrane lipid-mediated effects on GPCRs and (iii) GPCR oligomerization in a native-like membrane environment. We propose that, from the perspective of computational modeling, all three of these components need to be addressed in order to achieve a deeper understanding of GPCR functioning. Presently, we are able to simulate numerous lipid properties applying advanced computational techniques, although some barriers, such as the time-length of these simulations, need to be overcome. Implementing three-dimensional structures of GPCRs in such validated membrane systems can give novel insights in membrane-dependent receptor modulation and formation of higher order receptor complexes. Finally, more realistic GPCR-membrane models would provide a very useful tool in studying receptor behavior and its modulation by small drug-like ligands, a relevant issue for drug discovery.

Visualizing the Induced Binding of SH2-Phosphopeptide.

Approximately 100 proteins in the human genome contain an SH2 domain recognizing small flexible phosphopeptides. It is therefore important to understand in atomistic detail the way these peptides bind and the conformational changes that take place upon binding. Here, we obtained several spontaneous binding events between the p56 lck SH2 domain and the pYEEI peptide within 2 Å RMSD from the crystal structure and with kinetic rates compatible with experiments using high-throughput molecular dynamics simulations. Binding is achieved in two phases, fast contacts of the charged phospho-tyrosine and then rearrangement of the ligand involving the stabilization of two important loops in the SH2 domain. These observations provide insights into the binding pathways and induced conformations of the SH2-phosphopeptide complex which, due to the characteristics of SH2 domains, should be relevant for other SH2 recognition peptides.

Statistical analysis of global connectivity and activity distributions in cellular networks.

Various molecular interaction networks have been claimed to follow power-law decay for their global connectivity distribution. It has been proposed that there may be underlying generative models that explain this heavy-tailed behavior by self-reinforcement processes such as classical or hierarchical scale-free network models. Here, we analyze a comprehensive data set of protein-protein and transcriptional regulatory interaction networks in yeast, an Escherichia coli metabolic network, and gene activity profiles for different metabolic states in both organisms. We show that in all cases the networks have a heavy-tailed distribution, but most of them present significant differences from a power-law model according to a stringent statistical test. Those few data sets that have a statistically significant fit with a power-law model follow other distributions equally well. Thus, while our analysis supports that both global connectivity interaction networks and activity distributions are heavy-tailed, they are not generally described by any specific distribution model, leaving space for further inferences on generative models. Supplementary Material is available online at www.liebertonline.com.

An Implementation of the Smooth Particle Mesh Ewald Method on GPU Hardware.

The smooth particle mesh Ewald summation method is widely used to efficiently compute long-range electrostatic force terms in molecular dynamics simulations, and there has been considerable work in developing optimized implementations for a variety of parallel computer architectures. We describe an implementation for Nvidia graphical processing units (GPUs) which are general purpose computing devices with a high degree of intrinsic parallelism and arithmetic performance. We find that, for typical biomolecular simulations (e.g., DHFR, 26K atoms), a single GPU equipped workstation is able to provide sufficient performance to permit simulation rates of ≈50 ns/day when used in conjunction with the ACEMD molecular dynamics package (1) and exhibits an accuracy comparable to that of a reference double-precision CPU implementation.

Hybrid method coupling fluctuating hydrodynamics and molecular dynamics for the simulation of macromolecules.

We present a hybrid computational method for simulating the dynamics of macromolecules in solution which couples a mesoscale solver for the fluctuating hydrodynamics (FH) equations with molecular dynamics to describe the macromolecule. The two models interact through a dissipative Stokesian term first introduced by Ahlrichs and Dunweg [J. Chem. Phys. 111, 8225 (1999)]. We show that our method correctly captures the static and dynamical properties of polymer chains as predicted by the Zimm model. In particular, we show that the static conformations are best described when the ratio sigma/b=0.6, where sigma is the Lennard-Jones length parameter and b is the monomer bond length. We also find that the decay of the Rouse modes' autocorrelation function is better described with an analytical correction suggested by Ahlrichs and Dunweg. Our FH solver permits us to treat the fluid equation of state and transport parameters as direct simulation parameters. The expected independence of the chain dynamics on various choices of fluid equation of state and bulk viscosity is recovered, while excellent agreement is found for the temperature and shear viscosity dependence of center of mass diffusion between simulation results and predictions of the Zimm model. We find that Zimm model approximations start to fail when the Schmidt number Sc < or approximately 30. Finally, we investigate the importance of fluid fluctuations and show that using the preaveraged approximation for the hydrodynamic tensor leads to around 3% error in the diffusion coefficient for a polymer chain when the fluid discretization size is greater than 50 A.