Neuron splitting in compute-bound parallel network simulations enables runtime scaling with twice as many processors

Neuron tree topology equations can be split into two subtrees and solved on different processors with no change in accuracy, stability, or computational effort; communication costs involve only sending and receiving two double precision values by each subtree at each time step. Splitting cells is useful in attaining load balance in neural network simulations, especially when there is a wide range of cell sizes and the number of cells is about the same as the number of processors. For compute-bound simulations load balance results in almost ideal runtime scaling. Application of the cell splitting method to two published network models exhibits good runtime scaling on twice as many processors as could be effectively used with whole-cell balancing.     

Social networks of disease spread in the lower Illinois valley: a simulation approach

This study illustrates the use of disease modeling and simulation techniques to the study of the spread of disease within and between social networks. A Reed-Frost type model of disease spread is used to construct a simulation of the spread of tuberculosis within three prehistoric populations of the Lower Illinois River Valley during Middle Woodland, Late Woodland, and Mississippian times. A high and low population size was modeled for each time period. Late Woodland model 2 (low population estimate) is the only model that experienced pathogen extinction with host survival. The rest of the models experienced rapid and severe host population decline. The results of the simulation suggest that a social network size of between 180 and 440 persons is required under the conditions of this model for host-pathogen coexistence (i.e., endemicity) to occur. The severe population decline experienced by these populations suggests that tuberculosis as modeled here could not have existed in these populations. Future refinements of modeling and simulation techniques can provide additional insights into how disease spreads among social contacts.     

Parallel network simulations with NEURON

The NEURON simulation environment has been extended to support parallel network simulations. Each processor integrates the equations for its subnet over an interval equal to the minimum (interprocessor) presynaptic spike generation to postsynaptic spike delivery connection delay. The performance of three published network models with very different spike patterns exhibits superlinear speedup on Beowulf clusters and demonstrates that spike communication overhead is often less than the benefit of an increased fraction of the entire problem fitting into high speed cache. On the EPFL IBM Blue Gene, almost linear speedup was obtained up to 100 processors. Increasing one model from 500 to 40,000 realistic cells exhibited almost linear speedup on 2000 processors, with an integration time of 9.8 seconds and communication time of 1.3 seconds. The potential for speed-ups of several orders of magnitude makes practical the running of large network simulations that could otherwise not be explored. Action Editor: Alain Destexhe     

The hermeneutics of ecological simulation

Computer simulation has become important in ecological modeling, but there have been few assessments on how complex simulation models differ from more traditional analytic models. In Part I of this paper, I review the challenges faced in complex ecological modeling and how models have been used to gain theoretical purchase for understanding natural systems. I compare the use of traditional analytic simulation models and point how that the two methods require different kinds of practical engagement. I examine a case study of three models from the insect resistance literature in transgenic crops to illustrate and explore differences in analytic and computer simulation models. I argue that analyzing simulation models has been often inappropriately managed with expectations derived from handling analytic models. In Part II, I look at simulation as a hermeneutic practice. I argue that simulation models are a practice or techné. I the explore five aspects of philosophical hermeneutics that may be useful in complex ecological simulation: (1) an openness to multiple perspectives allowing multiple levels of scientific pluralism, (2) the hermeneutic circle, a back and forth in active communication among both modelers and ecologists; (3) the recognition of human factors and the nature of human practices as such, including recognizing the role of judgments and choices in the modeling enterprise; (4) the importance of play in modeling; (5) the non-closed nature of hermeneutic engagement, continued dialogue, and recognizing the situatedness, incompleteness, and tentative nature of simulation models.

A simulation of chopper neurons in the cochlear nucleus with wideband input from onset neurons

The unique temporal and spectral properties of chopper neurons in the cochlear nucleus cannot be fully explained by current popular models. A new model of sustained chopper neurons was therefore suggested based on the assumption that chopper neurons receive input both from onset neurons and the auditory nerve (Bahmer and Langner in Biol Cybern 95:4, 2006). As a result of the interaction of broadband input from onset neurons and narrowband input from the auditory nerve, the chopper neurons in our model are characterized by a remarkable combination of sharp frequency tuning to pure tones and faithful periodicity coding. Our simulations show that the width of the spectral integration of the onset neuron is crucial for both the precision of periodicity coding and their resolution of single components of sinusoidally amplitude-modulated sine waves. One may hypothesize, therefore, that it would be an advantage if the hearing system were able to adapt the spectral integration of onset neurons to varying stimulus conditions.     

Different types of rectification at electrical synapses made by a single crayfish neurone investigated experimentally and by computer simulation

Summary The rectification properties of electrical synapses made by the segmental giant (SG) neurone of crayfish (Pacifastacus leniusculus) were investigated. The SG acts as an interneurone, transmitting information from the giant command fibres (GFs) to the abdominal fast flexor (FF) motoneurones. The GF-SG (input) synapses are inwardly-rectifying electrical synapses, while the SG-FF (output) synapses are outwardly rectifying electrical synapses. This implies that a single neurone can make gap junction hemichannels with different rectification properties.The coupling coefficient of these synapses is dependent upon transjunctional potential. There is a standing gradient in resting potential between the GFs, SG and FFs, with the GFs the most hyperpolarized, and the FFs the most depolarized. The gradient thus biases each synapse into the low-conductance state under resting conditions.There is functional double rectification between the bilateral pairs of SGs within a single segment, such that depolarizing membrane potential changes of either SG pass to the other SG with less attenuation than do hyperpolarizing potential changes. Computer simulation suggests that this may result from coupling through the intermediary FF neurones.Abbreviations l left - r right - FF fast flexor motoneurone - GF giant fibre - LG lateral giant interneurone - MG medial giant interneurone - MoG motor giant motoneurone - R root, e.g. 1R1 is the first root on the left side - SG Segmental giant neurone     

Different ways to make neurons: parallel evolution in the SoxB family

Combining genome-wide analyses of binding sites and expression profiles generates a model for the functional evolution of two SOXB paralogous proteins in neurogenesis.     

Assessing the accuracy of subject-specific,muscle-model parameters determined by optimizing to match isometric strength

Biomechanical models are sensitive to the choice of model parameters. Therefore, determination of accurate subject specific model parameters is important. One approach to generate these parameters is to optimize the values such that the model output will match experimentally measured strength curves. This approach is attractive as it is inexpensive and should provide an excellent match to experimentally measured strength. However, given the problem of muscle redundancy, it is not clear that this approach generates accurate individual muscle forces. The purpose of this investigation is to evaluate this approach using simulated data to enable a direct comparison. It is hypothesized that the optimization approach will be able to recreate accurate muscle model parameters when information from measurable parameters is given. A model of isometric knee extension was developed to simulate a strength curve across a range of knee angles. In order to realistically recreate experimentally measured strength, random noise was added to the modeled strength. Parameters were solved for using a genetic search algorithm. When noise was added to the measurements the strength curve was reasonably recreated. However, the individual muscle model parameters and force curves were far less accurate. Based upon this examination, it is clear that very different sets of model parameters can recreate similar strength curves. Therefore, experimental variation in strength measurements has a significant influence on the results. Given the difficulty in accurately recreating individual muscle parameters, it may be more appropriate to perform simulations with lumped actuators representing similar muscles.     

Molecular dynamics simulation of a phospholipid membrane

We present the results of molecular dynamics (MD) simulations of a phospholipid membrane in water, including full atomic detail. The goal of the simulations was twofold: first we wanted to set up a simulation system which is able to reproduce experimental results and can serve as a model membrane in future simulations. This goal being reached it is then further possible to gain insight in to those properties that are experimentally more difficult to access. The system studied is dipalmitoylphosphatidylcholine/water, consisting of 5408 atoms. Using original force field parameters the membrane turned out to approach a gel-like state. With slight changes of the parameters, the system adopted a liquid-crystalline state. Separate 80 ps runs were performed on both the gel and liquid-crystalline systems. Comparison of MD results with reliable experimental data (bilayer repeat distance, surface area per lipid, tail order parameters, atom distributions) showed that our simulations, especially the one in the liquid-crystalline phase, can serve as a realistic model for a phospholipid membrane. Further analysis of the trajectories revealed valuable information on various properties. In the liquid-crystalline phase, the interface turns out to be quite diffuse, with water molecules penetrating into the bilayer to the position of the carbonyl groups. The 10–90% width of the interface turns out to be 1.3 nm and the width of the hydrocarbon interior 3.0 nm. The headgroup dipoles are oriented at a small angle with respect to the bilayer plane. The resulting charge distribution is almost completely cancelled by the water molecules. The electron density distribution shows a large dip in the middle of the membrane. In this part the tails are more flexible. The mean life time between dihedral transitions is 20 ps. The average number of gauche angles per tail is 3.5. The occurrence of kinks is not a significant feature.Abbreviations MD molecular dynamics - DPPC dipalmitoylphosphatidylcholine - SPC simple point charges - DPPE dipalmitoylphosphatidylethanolamine Correspondence to: H. J. C. Berendsen     

Computer simulation in template-directed oligonucleotide synthesis

Summary A computer simulation (KINSIM) modeling up to 33 competing reactions was used in order to investigate the product distribution in a template-directed oligonucleotide synthesis as a function of time and concentration of the reactants. The study is focused on the poly(C)-directed elongation reaction of an oligoguanylate (a 7-mer is chosen) with guanosine 5-monophosphate-2-methylimidazolide (2-MeImpG), the activated monomer. It is known that theelongation of oligoguanylates to form oligomeric products such as 8-mer, 9-mer, 10-mer, etc., is in competition with (1) thedimerization and further oligomerization reaction of 2-MeImpG that leads to the formation of dimers and short oligomers, and (2) thehydrolysis of 2-MeImpG that forms inactive guanosine 5-monophosphate, 5-GMP. Experimentally determined rate constants for the above three processes at 37°C and pH 7.95 were used in the simulation; the initial concentrations of 2-MeImpG, [M]_o, and of the oligoguanylate primer, [7-mer]_o, were varied, and KINSIM calculated the distribution of products as a function of time until equilibration was reached, i.e., when all the activated monomer has been consumed. In order to sort out how strongly the elongation reaction may be affected by the competing hydrolysis and dimerization, we also simulated the idealized situation in which these competing reactions do not occur. Simulation of the idealized system suggests that (1) the fraction of [7-mer]_o that has reacted as well as the product distribution after equilibration do not depend on the absolute concentrations of the reactants, but only on their ratio, [M]_o/[7-mer]_o; (2) the rate of elongation is proportional to [7-mer]_o and not to [M]_o; and (3) as the [M]_o/[7-mer]_o ratio increases longer oligomers are formed. The results of the computer simulation with the experimental system, i.e., elongation in the presence of both hydrolysis and dimerization, are similar to the ones obtained with the idealized system as long as dimerization and hydrolysis are not responsible for consuming a substantial fraction of 2-MeImpG.     

Dissection of a model for neuronal parabolic bursting

We have obtained new insight into the mechanisms for bursting in a class of theoretical models. We study Plant's model [24] for Aplysia R-15 to illustrate our view of these so-called parabolic bursters, which are characterized by low spike frequency at the beginning and end of a burst. By identifying and analyzing the fast and slow processes we show how they interact mutually to generate spike activity and the slow wave which underlies the burst pattern. Our treatment is essentially the first step of a singular perturbation approach presented from a geometrical viewpoint and carried out numerically with AUTO [12]. We determine the solution sets (steady state and oscillatory) of the fast subsystem with the slow variables treated as parameters. These solutions form the slow manifold over which the slow dynamics then define a burst trajectory. During the silent phase of a burst, the solution trajectory lies approximately on the steady state branch of the slow manifold and during the active phase of spiking, the trajectory sweeps through the oscillation branch. The parabolic nature of bursting arises from the (degenerate) homoclinic transition between the oscillatory branch and the steady state branch. We show that, for some parameter values, the trajectory remains strictly on the steady state branch (to produce a resting steady state or a pure slow wave without spike activity) or strictly in the oscillatory branch (continuous spike activity without silent phases). Plant's model has two slow variables: a calcium conductance and the intracellular free calcium concentration, which activates a potassium conductance. We also show how bursting arises from an alternative mechanism in which calcium inactivates the calcium current and the potassium conductance is insensitive to calcium. These and other biophysical interpretations are discussed.     

系统生物学中建模方法的研究现状及展望

系统生物学倡导利用系统论的思想和方法,从整体的高度分析、研究生命的复杂特性。这一点与实验生物学仅关注某一个或者某一些生物大分子是迥然不同的。系统生物学既要同时考虑多个层次、多种类型的生物信息,还要考虑时间因素。由于系统特性是由于不同组成部分、不同层次间相互作用而“涌现”出的新性质,因此,如果只是针对组成部分或单一层次的分析并不能真正准确地预测整体或高层次的行为。如何通过研究和整合去发现和理解“涌现”出的新的系统性质,是系统生物学面临的一个根本性的挑战。为了应对这一挑战,系统生物学,特别是计算系统生物学必须建立有效的方法,通过整合系统各个层次的信息,建立可反映该系统目前已知或已可测量的性质的物理、数学模型,并通过这样的模型来研究或预测目前还未知晓的系统性状。可以说：建模是系统生物学的最重要的研究手段之一。目前,生命科学的研究正逐步由对单一现象、单一过程的机械论式的描述型研究转向运用高通量实验技术获取海量生物信息,并在这些生物信息基础上建立物理、数学模型,最终通过建模与实验相接合的研究手段来定量阐述生命现象的本质规律。由于建模方法在系统生物学研究中的重要性,本文将对一些主要的建模类型,如定性建模方法;基于约束的建模方法;基于常微分/偏微分方程的定量建模和基于随机微分方程的定量建模方法等等分别予以简要介绍。     

Advances in defining signaling networks for the establishment of neuronal polarity

Neurons are highly polarized cells that have structurally and functionally distinct processes called axons and dendrites. How neurons establish polarity is one of the fundamental questions of neuroscience. In the last decade, significant progress has been made in identifying and understanding the molecular mechanisms responsible for neuronal polarization, primarily through researches conducted on cultured neurons. Advances in phosphoproteomics technologies and molecular tools have enabled comprehensive signal analysis and visualization and manipulation of signaling molecules for analyzing neuronal polarity. Furthermore, advances in gene transfer techniques have revealed the role of extracellular and intracellular signaling molecules in neuronal polarization in vivo. This review discusses the latest insights and techniques for the elucidation of the molecular mechanisms that control neuronal polarity.     

Multiscale simulation for thermo-hydrodynamic lubrication of a polymeric liquid between parallel plates

Thermo-hydrodynamic lubrication of a polymeric liquid composed of short chains between parallel plates is analysed by a multi-scale simulation, i.e. the synchronised molecular dynamics simulation via macroscopic heat and momentum transfer, which has been recently developed by us. The rheological properties and conformation of polymer chains coupled with the temperature rise caused by local viscous heating are investigated with a non-dimensional parameter, i.e. the Nahme–Griffith number, which is defined by the ratio of the viscous heating to the thermal conduction at the characteristic temperature required to sufficiently change the viscosity. The present simulation demonstrates that strong shear thinning and transitional behaviour of the conformation of the polymer chains occurs with a rapid temperature rise when the Nahme–Griffith number exceeds unity.     

A Model of Human Muscle Energy Expenditure

A model of muscle energy expenditure was developed for predicting thermal, as well as mechanical energy liberation during simulated muscle contractions. The model was designed to yield energy (heat and work) rate predictions appropriate for human skeletal muscle contracting at normal body temperature. The basic form of the present model is similar to many previous models of muscle energy expenditure, but parameter values were based almost entirely on mammalian muscle data, with preference given to human data where possible. Nonlinear phenomena associated with submaximal activation were also incorporated. The muscle energy model was evaluated at varying levels of complexity, ranging from simulated contractions of isolated muscle, to simulations of whole body locomotion. In all cases, acceptable agreement was found between simulated and experimental energy liberation. The present model should be useful in future studies of the energetics of human movement using forward dynamic computer simulation.     

Microfluidic formation of single cell array for parallel analysis of Ca release-activated Ca (CRAC) channel activation and inhibition

High-throughput single cell analysis is required for understanding and predicting the complex stochastic responses of individual cells in changing environments. We have designed a microfluidic device consisting of parallel, independent channels with cell-docking structures for the formation of an array of individual cells. The microfluidic cell array was used to quantify single cell responses and the distribution of response patterns of calcium channels among a population of individual cells. In this device, 15 cell-docking units in each channel were fabricated with each unit containing 5 sandbag structures, such that an array of individual cells was formed in 8 independent channels. Single cell responses to different treatments in different channels were monitored in parallel to study the effects of the specific activator and inhibitor of the Ca²⁺ release-activated Ca²⁺ (CRAC) channels. Multichannel detection was performed to obtain the response patterns of the population of cells within this single cell array. The results demonstrate that it is possible to acquire single cell features in multichannels simultaneously with passive structural control, which provides an opportunity for high-throughput single cell response analysis in a microfluidic chip.     

Contact pressure models for spiral phyllotaxis and their computer simulation

We present a simple, biologically motivated model for the creation of phyllotactical patterns in capitula and their computer simulation. An in-depth investigation of Ridley's contact pressure model is performed and a refinement of contact pressure between primordia on the basis of local centroidal Voronoi relaxation is presented. Using this method in combination with Hofmeisters rule for placing new primordia creates stable patterns with Fibonacci spirals of high degree for a large range of initial conditions.     

Reliability of regulatory networks and its evolution

The problem of reliability of the dynamics in biological regulatory networks is studied in the framework of a generalized Boolean network model with continuous timing and noise. Using well-known artificial genetic networks such as the repressilator, we discuss concepts of reliability of rhythmic attractors. In a simple evolution process we investigate how overall network structure affects the reliability of the dynamics. In the course of the evolution, networks are selected for reliable dynamics. We find that most networks can be easily evolved towards reliable functioning while preserving the original function.     

树干径流过程的计算机模拟

本文应用系统分析方法,将树干径流过程这一连续系统离散化,编制程序,以计算机直接模拟过程行为,绕过了解微分方程之烦恼,并取得了满意的结果。此举对扩展系统分析概念,研究连续系统颇有意义。     

Growth and refuge location in continuous,modular organisms: experimental and computer simulation studies

Summary Guilds of crustose bryozoans exhibited a range of growth outlines, varying from circular to elongate, and the stones on which they grew contained a range of refuge types, varying in size from a few mm² to a few hundred cm². These refuges, if encountered by the growing colonies of competitively subordinate species such as Electra pilosa L., enabled the latter to avoid competitive exclusion due to overgrowth mortality. Computer simulations suggested that, although modular organisms which are elongate are, in general, more likely to encounter spatial refuges than those with a different shape through vegetative growth, shape is less important than the size of the organisms themselves or the size and density of the spatial refuges. Moreover significant interactions between these 4 variables showed that they should be considered together and not in isolation from one another. Examples are given of variable growth forms in continuous, modular organisms from different taxa, including plants, and the importance of these predictive studies to refuge location by modular organisms is discussed.