Structure Prediction Modeling Research Articles

Structured prediction models for RNN based sequence labeling in clinical text.

Sequence labeling is a widely used method for named entity recognition and information extraction from unstructured natural language data. In clinical domain one major application of sequence labeling involves extraction of medical entities such as medication, indication, and side-effects from Electronic Health Record narratives. Sequence labeling in this domain, presents its own set of challenges and objectives. In this work we experimented with various CRF based structured learning models with Recurrent Neural Networks. We extend the previously studied LSTM-CRF models with explicit modeling of pairwise potentials. We also propose an approximate version of skip-chain CRF inference with RNN potentials. We use these methodologies for structured prediction in order to improve the exact phrase detection of various medical entities.

Whole genome annotation and comparative genomic analyses of bio-control fungus Purpureocillium lilacinum.

BackgroundThe fungus Purpureocillium lilacinum is widely known as a biological control agent against plant parasitic nematodes. This research article consists of genomic annotation of the first draft of whole genome sequence of P. lilacinum. The study aims to decipher the putative genetic components of the fungus involved in nematode pathogenesis by performing comparative genomic analysis with nine closely related fungal species in Hypocreales.Resultsde novo genomic assembly was done and a total of 301 scaffolds were constructed for P. lilacinum genomic DNA. By employing structural genome prediction models, 13, 266 genes coding for proteins were predicted in the genome. Approximately 73 % of the predicted genes were functionally annotated using Blastp, InterProScan and Gene Ontology. A 14.7 % fraction of the predicted genes shared significant homology with genes in the Pathogen Host Interactions (PHI) database. The phylogenomic analysis carried out using maximum likelihood RAxML algorithm provided insight into the evolutionary relationship of P. lilacinum. In congruence with other closely related species in the Hypocreales namely, Metarhizium spp., Pochonia chlamydosporia, Cordyceps militaris, Trichoderma reesei and Fusarium spp., P. lilacinum has large gene sets coding for G-protein coupled receptors (GPCRs), proteases, glycoside hydrolases and carbohydrate esterases that are required for degradation of nematode-egg shell components. Screening of the genome by Antibiotics & Secondary Metabolite Analysis Shell (AntiSMASH) pipeline indicated that the genome potentially codes for a variety of secondary metabolites, possibly required for adaptation to heterogeneous lifestyles reported for P. lilacinum. Significant up-regulation of subtilisin-like serine protease genes in presence of nematode eggs in quantitative real-time analyses suggested potential role of serine proteases in nematode pathogenesis.ConclusionsThe data offer a better understanding of Purpureocillium lilacinum genome and will enhance our understanding on the molecular mechanism involved in nematophagy.Electronic supplementary materialThe online version of this article (doi:10.1186/s12864-015-2229-2) contains supplementary material, which is available to authorized users.

Informing phenomenological structural bone remodelling with a mechanistic poroelastic model.

Studies suggest that fluid motion in the extracellular space may be involved in the cellular mechanosensitivity at play in the bone tissue adaptation process. Previously, the authors developed a mesoscale predictive structural model of the femur using truss elements to represent trabecular bone, relying on a phenomenological strain-based bone adaptation algorithm. In order to introduce a response to bending and shear, the authors considered the use of beam elements, requiring a new formulation of the bone adaptation drivers. The primary goal of the study presented here was to isolate phenomenological drivers based on the results of a mechanistic approach to be used with a beam element representation of trabecular bone in mesoscale structural modelling. A single-beam model and a microscale poroelastic model of a single trabecula were developed. A mechanistic iterative adaptation algorithm was implemented based on fluid motion velocity through the bone matrix pores to predict the remodelled geometries of the poroelastic trabecula under 42 different loading scenarios. Regression analyses were used to correlate the changes in poroelastic trabecula thickness and orientation to the initial strain outputs of the beam model. Linear (R^2>0.998) and third-order polynomial (R^2 >0.98) relationships were found between change in cross section and axial strain at the central axis, and between beam reorientation and ratio of bending strain to axial strain, respectively. Implementing these relationships into the phenomenological predictive algorithm for the mesoscale structural femur has the potential to produce a model combining biofidelic structure and mechanical behaviour with computational efficiency.

Prediction of the structures of the plant-specific regions of vascular plant cellulose synthases and correlated functional analysis

Seed plants express cellulose synthase (CESA) protein isoforms with non-redundant functions, but how the isoforms function differently is unknown. Compared to bacterial cellulose synthases, CESAs have two insertions in the large cytosolic loop: the relatively well-conserved Plant Conserved Region (P-CR) and a Class Specific Region (CSR) that varies between CESAs. Absent any atomic structure of a plant CESA, we used ab initio protein structure prediction and molecular modeling to explore how these plant-specific regions may modulate CESA function. We modeled P-CR and CSR peptides from Arabidopsis thaliana CESAs representing the six clades of seed plant CESAs. As expected, the predicted wild type P-CR structures were similar. Modeling of the mutant P-CR of Atcesa8 R362K (fra6) suggested that changes in local structural stability and surface electrostatics may cause the mutant phenotype. Among CSRs within CESAs required for primary wall cellulose synthesis, the amino sequence and the modeled arrangement of helices was most similar in AtCESA1 and AtCESA3. Genetic complementation of known Arabidopsis mutants showed that the CSRs of AtCESA1 and AtCESA3 can function interchangeably in vivo. Analysis of protein surface electrostatics led to ideas about how the surface charges on CSRs may mediate protein–protein interactions. Refined modeling of the P-CR and CSR regions of GhCESA1 from cotton modified their tertiary structures, spatial relationships to the catalytic domain, and preliminary predictions about CESA oligomer formation. Cumulatively, the results provide structural clues about the function of plant-specific regions of CESA.

A predictive structural model for bulk metallic glasses.

Great progress has been made in understanding the atomic structure of metallic glasses, but there is still no clear connection between atomic structure and glass-forming ability. Here we give new insights into perhaps the most important question in the field of amorphous metals: how can glass-forming ability be predicted from atomic structure? We give a new approach to modelling metallic glass atomic structures by solving three long-standing problems: we discover a new family of structural defects that discourage glass formation; we impose efficient local packing around all atoms simultaneously; and we enforce structural self-consistency. Fewer than a dozen binary structures satisfy these constraints, but extra degrees of freedom in structures with three or more different atom sizes significantly expand the number of relatively stable, ‘bulk' metallic glasses. The present work gives a new approach towards achieving the long-sought goal of a predictive capability for bulk metallic glasses.

A Meta-heuristic Framework for Secondary Protein Structure Prediction using BAT-FLANN Optimization Algorithm

Background/Objectives: Proteins are the fundamental units of biology; the mechanism by which primary sequence of proteins is predicted into its secondary structure is not yet accurately achieved. Methods/Statistical analysis: In this paper, BAT inspired FLANN (Functional Link Artificial Neural Network) model for protein secondary structure prediction with low computation cost and accuracy has been proposed. The proposed model consists of three different phases; i) First, the primary sequence of amino acid is converted into dynamic matrix for different window sizes then this dynamic matrix is used to derive correlation matrix, ii) Second, FLANN is used to classify each sequence of correlation matrix with different learning parameters and random weights. BAT inspired optimization algorithm has been used to optimize the weight and learning parameters of BAT-FLANN, and (iii) finally, refinement of secondary structure result. Results: Experiments were conducted with real datasets of some primary sequence on RS126 and CB396 datasets. Proposed method has been compared with existing DSC, NNSSP, PHD, PREDATOR, ZPRED, MULPRED, SVM models and found to be more promising. Conclusion/Application: The proposed method achieves average Q3 accuracy 81.2% and 82.7% for CB396 and RS126 dataset respectively. Moreover the segment overlap (SOV) is 76.1% and 75.3% for CB396 and RS126 dataset respectively.

An integer programming model for protein structure prediction using the 3D-HP side chain model

In spite of the fact that many simplified model variants of protein structure prediction have been widely studied in the past years, few attention has been given to discrete models with side chains, for which there is no specific benchmark. In this paper, we propose an integer programming model for the 3D-HP side chain protein structure prediction problem. The model accounts for the energy resulting from all types of interactions, between pairs of backbone elements, hydrophilic side chains and hydrophobic side chains. Three sets of instances, modified from the literature, were used in the experiments, and the maximum number of non-local hydrophobic contact was found using the ILOG CPLEX optimization package. We offer the optimal solution found for several instances of the benchmark. It is expected that the mathematical model allow further studies of the protein structure prediction with side chains and may, for some cases, provide new optimal values or new bounds that would rekindle the interest to this fascinating problem domain.

GPstruct: Bayesian Structured Prediction Using Gaussian Processes.

We introduce a conceptually novel structured prediction model, GPstruct, which is kernelized, non-parametric and Bayesian, by design. We motivate the model with respect to existing approaches, among others, conditional random fields (CRFs), maximum margin Markov networks (M3N), and structured support vector machines (SVMstruct), which embody only a subset of its properties. We present an inference procedure based on Markov Chain Monte Carlo. The framework can be instantiated for a wide range of structured objects such as linear chains, trees, grids, and other general graphs. As a proof of concept, the model is benchmarked on several natural language processing tasks and a video gesture segmentation task involving a linear chain structure. We show prediction accuracies for GPstruct which are comparable to or exceeding those of CRFs and SVMstruct.

Coarse-Grained HiRE-RNA Model for ab Initio RNA Folding beyond Simple Molecules, Including Noncanonical and Multiple Base Pairings

HiRE-RNA is a coarse-grained model for RNA structure prediction and the dynamical study of RNA folding. Using a reduced set of particles and detailed interactions accounting for base-pairing and stacking, we show that noncanonical and multiple base interactions are necessary to capture the full physical behavior of complex RNAs. In this paper, we give a full account of the model and present results on the folding, stability, and free energy surfaces of 16 systems with 12 to 76 nucleotides of increasingly complex architectures, ranging from monomers to dimers, using a total of 850 μs of simulation time.

I-TASSER server: new development for protein structure and function predictions.

The I-TASSER server (http://zhanglab.ccmb.med.umich.edu/I-TASSER) is an online resource for automated protein structure prediction and structure-based function annotation. In I-TASSER, structural templates are first recognized from the PDB using multiple threading alignment approaches. Full-length structure models are then constructed by iterative fragment assembly simulations. The functional insights are finally derived by matching the predicted structure models with known proteins in the function databases. Although the server has been widely used for various biological and biomedical investigations, numerous comments and suggestions have been reported from the user community. In this article, we summarize recent developments on the I-TASSER server, which were designed to address the requirements from the user community and to increase the accuracy of modeling predictions. Focuses have been made on the introduction of new methods for atomic-level structure refinement, local structure quality estimation and biological function annotations. We expect that these new developments will improve the quality of the I-TASSER server and further facilitate its use by the community for high-resolution structure and function prediction.

IntFOLD: an integrated server for modelling protein structures and functions from amino acid sequences.

IntFOLD is an independent web server that integrates our leading methods for structure and function prediction. The server provides a simple unified interface that aims to make complex protein modelling data more accessible to life scientists. The server web interface is designed to be intuitive and integrates a complex set of quantitative data, so that 3D modelling results can be viewed on a single page and interpreted by non-expert modellers at a glance. The only required input to the server is an amino acid sequence for the target protein. Here we describe major performance and user interface updates to the server, which comprises an integrated pipeline of methods for: tertiary structure prediction, global and local 3D model quality assessment, disorder prediction, structural domain prediction, function prediction and modelling of protein-ligand interactions. The server has been independently validated during numerous CASP (Critical Assessment of Techniques for Protein Structure Prediction) experiments, as well as being continuously evaluated by the CAMEO (Continuous Automated Model Evaluation) project. The IntFOLD server is available at: http://www.reading.ac.uk/bioinf/IntFOLD/

Structural Learning with Amortized Inference

Training a structured prediction model involves performing several loss-augmented inference steps. Over the lifetime of the training, many of these inference problems, although different, share the same solution. We propose AI-DCD, an Amortized Inference framework for Dual Coordinate Descent method, an approximate learning algorithm, that accelerates the training process by exploiting this redundancy of solutions, without compromising the performance of the model. We show the efficacy of our method by training a structured SVM using dual coordinate descent for an entityrelation extraction task. Our method learns the same model as an exact training algorithm would, but call the inference engine only in 10% – 24% of the inference problems encountered during training. We observe similar gains on a multi-label classification task and with a Structured Perceptron model for the entity-relation task.

Fabrication of 3D nanostructures by multidirectional UV lithography and predictive structural modeling

This paper presents the fabrication and modeling of three-dimensional (3D) nanostructures by automated multidirectional ultraviolet (UV) lithography, which is a fast, cost-effective, manufacturable fabrication method. Multidirectional UV exposure is performed using a static UV light source equipped with a tilt-rotational substrate holder. A glass substrate with a nanopatterned chrome layer is utilized as both a photomask and a substrate, for which a backside UV exposure scheme is used. For the analytical modeling of the shape of fabricated nanostructures, UV exposure dosage, diffraction and refraction effects, and absorption rate are taken into account. For more accurate process predictive models, a commercially available multiphysics simulation tool is used. The structural shapes predicted from analytical calculation and simulation are compared with the fabricated ones for which various 3D nanoscale test structures are fabricated such as an inclined nanopillar array and a vertical triangular slab. Also, nanostructures with multiple heights are successfully implemented from single layer photoresist by controlling the UV exposure dosage and tilt angles. A tripod embedded horn and a triangular-slab embedded horn are demonstrated.

A Nucleobase-Centric Coarse-Grained Model for Structure Prediction of RNA Fragments

The structural characterization of RNA usually poses additional challenges when compared to other biomolecular systems. For example, there is a relatively scarce amount of structural data available, which demands the development of three-dimensional structure prediction tools. In addition, full-resolution simulations can be a hard task not only because of the complexity of the interactions involved, but also due to the limitations of the current force fields. At this point, coarse-grained simulations are a good candidate to fill the gaps in this growing research field. The use of these techniques has rapidly increased in the past decades in the study of biological systems on which the experiments require an additional interpretation or where an atomistic computational approach results difficult or unfeasible. Nevertheless, the development of coarse-grained models involves the understanding of the main structural features of the original system, which can represent a challenge by itself.In this work, we present a knowledge-base coarse-grained model for RNA structure prediction, representing each nucleotide by a single anisotropic particle. The mapping and the main interactions are designed to reproduce the geometrical distribution of the closest pairs of nucleotides, extracted from a set of ribosomal structures. The model is inspired in the ESCORE function [1], a knowledge-based scoring function that has been shown to perform better compared to fully atomistic techniques in identifying native-like structures from a set of decoys. Its minimalistic nature and successful application on a broad range of structures straightforwardly suggest a representation for a coarse-grained approach. We show the preliminary results of our simulations and discuss the role of pair interactions in the prediction of RNA structures.[1] S. Bottaro, F. Di Palma and G. Bussi, “The Role of Nucleobase Interactions in RNA Structure and Dynamics”, Nucleic Acids Res., accepted.

Bank Default Prediction: A Comparative Model using Principal Component Analysis

Bank default prediction continues to draw attention given the ongoing effects of the recent financial crisis. Seminal works have found that structural models are better predictors of default. In this paper I argue that accounting models predictive ability have been weakened due to the multicollinearity problem and propose principal component analysis to improve the accounting model. The paper then compares accounting and structural default prediction models using a logit analysis and further evaluates the performance of a combination of accounting and structural default models to predict default. The paper uses panel data on US banks from the Federal Deposit Insurance Corporation database between 1995-2012 and the analysis is developed on 519 defaulted bank years and 5,965 non defaulted bank years. The accounting model is improved and outperforms the structural model; the study also finds that a combination of both models performs better than any one model at predicting default in the US banking system.

An iterative hybrid random-interval structural reliability analysis

An iterative hybrid structural dynamic reliability prediction model has been developed under multiple-time interval loads with and without consideration of stochastic structural strength degradation. Firstly, multiple-time interval loads have been substituted by the equivalent interval load. The equivalent interval load and structural strength are assumed as random variables. For structural reliability problem with random and interval variables, the interval variables can be converted to uniformly distributed random variables. Secondly, structural reliability with interval and stochastic variables is computed iteratively using the first order second moment method according to the stress-strength interference theory. Finally, the proposed method is verified by three examples which show that the method is practicable, rational and gives accurate prediction.

Predicted structure model of Bungarotoxin from Bungarus fasciatus snake.

Snake venoms are cocktails comprising combinations of different proteins, peptides, enzymes and toxins. Snake toxins have diverse characteristics having different molecular configuration, structure and mode of action. Many toxins derived from snake venom have distinct pharmacological activities. Venom from Bungarus fasciatus (commonly known as banded krait) is a species of elapid snake found on the South East Asia and Indian sub-continent, mainly contains neurotoxins. Beta bungartotoxin is the major fraction of Bungarus venom and particularly act pre-synaptically by obstructing neurotransmitter release. This toxin in other snake species functionally forms a heterodimer containing two different subunits (A and B). Dimerization of these two chains is a pre-requisite for the proper functionality of this protein. However, B. fasciatus bungartotoxin contains only B chain and their structural orientation in yet to be resolved. Therefore, it is of interest to describe the predicted structure model of the toxin for functional insights. In this work we analyzed the neurotoxic nature, their alignments, secondary and three dimensional structures, functions, active sites and stability with the help of different bioinformatical tools. A comprehensive analysis of the predicted model provides approaching to the functional interpretation of its molecular action.

The Impact of Equity Misvaluation on PredictiveAccuracy of Bankruptcy Models

This article examines the impact of equity misvaluation on the predictive accuracy of bankruptcy models. The authors find that structural bankruptcy prediction models are not affected by misvaluation. For hazard models, however, forecasting accuracy for properly valued firms is greater than for misvalued firms, and model forecasting accuracy improves significantly if model coefficients vary with misvaluation. The results show the importance of taking stock market misvaluation into account when forecasting bankruptcies using hazard models.

Constraint-handling through multi-objective optimization: The hydrophobic-polar model for protein structure prediction

In the multi-objective approach to constraint-handling, a constrained problem is transformed into an unconstrained one by defining additional optimization criteria to account for the problem constraints. In this paper, this approach is explored in the context of the hydrophobic-polar model, a simplified yet challenging representation of the protein structure prediction problem. Although focused on such a particular case of study, this research work is intended to contribute to the general understanding of the multi-objective constraint-handling strategy. First, a detailed analysis was conducted to investigate the extent to which this strategy impacts on the characteristics of the fitness landscape. As a result, it was found that an important fraction of the infeasibility translates into neutrality. This neutrality defines potentially shorter paths to move through the landscape, which can also be exploited to escape from local optima. By studying different mechanisms, the second part of this work highlights the relevance of introducing a proper search bias when handling constraints by multi-objective optimization. Finally, the suitability of the multi-objective approach was further evaluated in terms of its ability to effectively guide the search process. This strategy significantly improved the performance of the considered search algorithms when compared with respect to commonly adopted techniques from the literature.

Learning To Fold Proteins Using Energy Landscape Theory.

This review is a tutorial for scientists interested in the problem of protein structure prediction, particularly those interested in using coarse-grained molecular dynamics models that are optimized using lessons learned from the energy landscape theory of protein folding. We also present a review of the results of the AMH/AMC/AMW/AWSEM family of coarse-grained molecular dynamics protein folding models to illustrate the points covered in the first part of the article. Accurate coarse-grained structure prediction models can be used to investigate a wide range of conceptual and mechanistic issues outside of protein structure prediction; specifically, the paper concludes by reviewing how AWSEM has in recent years been able to elucidate questions related to the unusual kinetic behavior of artificially designed proteins, multidomain protein misfolding, and the initial stages of protein aggregation.