Splitting statistical potentials into meaningful scoring functions: testing the prediction of near-native structures from decoy conformations

Background: Recent advances on high-throughput technologies have produced a vast amount of protein sequences, while the number of high-resolution structures has seen a limited increase. This has impelled the production of many strategies to built protein structures from its sequence, generating a considerable amount of alternative models. The selection of the closest model to the native conformation has thus become crucial for structure prediction. Several methods have been developed to score protein models by energies, knowledge-based potentials and combination of both.Results: Here, we present and demonstrate a theory to split the knowledge-based potentials in scoring terms biologically meaningful and to combine them in new scores to predict near-native structures. Our strategy allows circumventing the problem of defining the reference state. In this approach we give the proof for a simple and linear application that can be further improved by optimizing the combination of Zscores. Using the simplest composite score () we obtained predictions similar to state-of-the-art methods. Besides, our approach has the advantage of identifying the most relevant terms involved in the stability of the protein structure. Finally, we also use the composite Zscores to assess the conformation of models and to detect local errors.Conclusion: We have introduced a method to split knowledge-based potentials and to solve the problem of defining a reference state. The new scores have detected near-native structures as accurately as state-of-art methods and have been successful to identify wrongly modeled regions of many near-native conformations.

Toward on-the-fly quantum mechanical/molecular mechanical (QM/MM) docking: development and benchmark of a scoring function.

We address the challenges of treating polarization and covalent interactions in docking by developing a hybrid quantum mechanical/molecular mechanical (QM/MM) scoring function based on the semiempirical self-consistent charge density functional tight-binding (SCC-DFTB) method and the CHARMM force field. To benchmark this scoring function within the EADock DSS docking algorithm, we created a publicly available dataset of high-quality X-ray structures of zinc metalloproteins ( http://www.molecular-modelling.ch/resources.php ). For zinc-bound ligands (226 complexes), the QM/MM scoring yielded a substantially improved success rate compared to the classical scoring function (77.0% vs 61.5%), while, for allosteric ligands (55 complexes), the success rate remained constant (49.1%). The QM/MM scoring significantly improved the detection of correct zinc-binding geometries and improved the docking success rate by more than 20% for several important drug targets. The performance of both the classical and the QM/MM scoring functions compare favorably to the performance of AutoDock4, AutoDock4Zn, and AutoDock Vina.

Assembling genes from predicted exons in linear time with dynamic programming

In a number of programs for gene structure prediction in higher eukaryotic genomic sequences, exon prediction is decoupled from gene assembly: a large pool of candidate exons is predicted and scored from features located in the query DNA sequence, and candidate genes are assembled from such a pool as sequences of nonoverlapping frame-compatible exons. Genes are scored as a function of the scores of the assembled exons, and the highest scoring candidate gene is assumed to be the most likely gene encoded by the query DNA sequence. Considering additive gene scoring functions, currently available algorithms to determine such a highest scoring candidate gene run in time proportional to the square of the number of predicted exons. Here, we present an algorithm whose running time grows only linearly with the size of the set of predicted exons. Polynomial algorithms rely on the fact that, while scanning the set of predicted exons, the highest scoring gene ending in a given exon can be obtained by appending the exon to the highest scoring among the highest scoring genes ending at each compatible preceding exon. The algorithm here relies on the simple fact that such highest scoring gene can be stored and updated. This requires scanning the set of predicted exons simultaneously by increasing acceptor and donor position. On the other hand, the algorithm described here does not assume an underlying gene structure model. Indeed, the definition of valid gene structures is externally defined in the so-called Gene Model. The Gene Model specifies simply which gene features are allowed immediately upstream which other gene features in valid gene structures. This allows for great flexibility in formulating the gene identification problem. In particular it allows for multiple-gene two-strand predictions and for considering gene features other than coding exons (such as promoter elements) in valid gene structures.

Étude des interactions supramoléculaires par modélisation moléculaire

L’avancée des infrastructures informatiques a permis l’émergence de la modélisation moléculaire. À cet effet, une multitude de modèles mathématiques sont aujourd’hui disponibles pour simuler différents systèmes chimiques. À l’aide de la modélisation moléculaire, différents types d’interactions chimiques ont été observés. À partir des systèmes les plus simples permettant l’utilisation de modèles quantiques rigoureux, une série d’approximations a été considérée pour rendre envisageable la simulation de systèmes moléculaires de plus en plus complexes. En premier lieu, la théorie de la fonctionnelle de densité dépendante du temps a été utilisée pour simuler les énergies d’excitation de molécules photoactives. De manière similaire, la DFT indépendante du temps a permis la simulation du pont hydrogène intramoléculaire de structures analogues au 1,3,5-triazapentadiène et la rationalisation de la stabilité des états de transition. Par la suite, la dynamique moléculaire et la mécanique moléculaire ont permis de simuler les interactions d’un trimère d’acide cholique et d’un pyrène dans différents solvants. Cette même méthodologie a été utilisée pour simuler les interactions d’un rotaxane-parapluie à l’interface d’un système biphasique. Finalement, l’arrimage moléculaire et les fonctions de score ont été utilisés pour simuler les interactions intermoléculaires entre une protéine et des milliers de candidats moléculaires. Les résultats ont permis de mettre en place une stratégie de développement d’un nouvel inhibiteur enzymatique.

Une nouvelle approche computationnelle pour la découverte des sites de fixation de facteurs de transcription à l’ADN, adaptée aux données de ChIP-chip et de ChIP-séquençage

Les facteurs de transcription sont des protéines spécialisées qui jouent un rôle important dans différents processus biologiques tel que la différenciation, le cycle cellulaire et la tumorigenèse. Ils régulent la transcription des gènes en se fixant sur des séquences d’ADN spécifiques (éléments cis-régulateurs). L’identification de ces éléments est une étape cruciale dans la compréhension des réseaux de régulation des gènes. Avec l’avènement des technologies de séquençage à haut débit, l’identification de tout les éléments fonctionnels dans les génomes, incluant gènes et éléments cis-régulateurs a connu une avancée considérable. Alors qu’on est arrivé à estimer le nombre de gènes chez différentes espèces, l’information sur les éléments qui contrôlent et orchestrent la régulation de ces gènes est encore mal définie. Grace aux techniques de ChIP-chip et de ChIP-séquençage il est possible d’identifier toutes les régions du génome qui sont liées par un facteur de transcription d’intérêt. Plusieurs approches computationnelles ont été développées pour prédire les sites fixés par les facteurs de transcription. Ces approches sont classées en deux catégories principales: les algorithmes énumératifs et probabilistes. Toutefois, plusieurs études ont montré que ces approches génèrent des taux élevés de faux négatifs et de faux positifs ce qui rend difficile l’interprétation des résultats et par conséquent leur validation expérimentale. Dans cette thèse, nous avons ciblé deux objectifs. Le premier objectif a été de développer une nouvelle approche pour la découverte des sites de fixation des facteurs de transcription à l’ADN (SAMD-ChIP) adaptée aux données de ChIP-chip et de ChIP-séquençage. Notre approche implémente un algorithme hybride qui combine les deux stratégies énumérative et probabiliste, afin d’exploiter les performances de chacune d’entre elles. Notre approche a montré ses performances, comparée aux outils de découvertes de motifs existants sur des jeux de données simulées et des jeux de données de ChIP-chip et de ChIP-séquençage. SAMD-ChIP présente aussi l’avantage d’exploiter les propriétés de distributions des sites liés par les facteurs de transcription autour du centre des régions liées afin de limiter la prédiction aux motifs qui sont enrichis dans une fenêtre de longueur fixe autour du centre de ces régions. Les facteurs de transcription agissent rarement seuls. Ils forment souvent des complexes pour interagir avec l’ADN pour réguler leurs gènes cibles. Ces interactions impliquent des facteurs de transcription dont les sites de fixation à l’ADN sont localisés proches les uns des autres ou bien médier par des boucles de chromatine. Notre deuxième objectif a été d’exploiter la proximité spatiale des sites liés par les facteurs de transcription dans les régions de ChIP-chip et de ChIP-séquençage pour développer une approche pour la prédiction des motifs composites (motifs composés par deux sites et séparés par un espacement de taille fixe). Nous avons testé ce module pour prédire la co-localisation entre les deux demi-sites ERE qui forment le site ERE, lié par le récepteur des œstrogènes ERα. Ce module a été incorporé à notre outil de découverte de motifs SAMD-ChIP.

Docking and molecular dynamics simulation of quinone compounds with trypanocidal activity

In this work, two different docking programs were used, AutoDock and FlexX, which use different types of scoring functions and searching methods. The docking poses of all quinone compounds studied stayed in the same region in the trypanothione reductase. This region is a hydrophobic pocket near to Phe396, Pro398 and Leu399 amino acid residues. The compounds studied displays a higher affinity in trypanothione reductase (TR) than glutathione reductase (GR), since only two out of 28 quinone compounds presented more favorable docking energy in the site of human enzyme. The interaction of quinone compounds with the TR enzyme is in agreement with other studies, which showed different binding sites from the ones formed by cysteines 52 and 58. To verify the results obtained by docking, we carried out a molecular dynamics simulation with the compounds that presented the highest and lowest docking energies. The results showed that the root mean square deviation (RMSD) between the initial and final pose were very small. In addition, the hydrogen bond pattern was conserved along the simulation. In the parasite enzyme, the amino acid residues Leu399, Met400 and Lys402 are replaced in the human enzyme by Met406, Tyr407 and Ala409, respectively. In view of the fact that Leu399 is an amino acid of the Z site, this difference could be explored to design selective inhibitors of TR.

Rainfall spatial predictions: a two-part model and its assessment

Spatial prediction of hourly rainfall via radar calibration is addressed. The change of support problem (COSP), arising when the spatial supports of different data sources do not coincide, is faced in a non-Gaussian setting; in fact, hourly rainfall in Emilia-Romagna region, in Italy, is characterized by abundance of zero values and right-skeweness of the distribution of positive amounts. Rain gauge direct measurements on sparsely distributed locations and hourly cumulated radar grids are provided by the ARPA-SIMC Emilia-Romagna. We propose a three-stage Bayesian hierarchical model for radar calibration, exploiting rain gauges as reference measure. Rain probability and amounts are modeled via linear relationships with radar in the log scale; spatial correlated Gaussian effects capture the residual information. We employ a probit link for rainfall probability and Gamma distribution for rainfall positive amounts; the two steps are joined via a two-part semicontinuous model. Three model specifications differently addressing COSP are presented; in particular, a stochastic weighting of all radar pixels, driven by a latent Gaussian process defined on the grid, is employed. Estimation is performed via MCMC procedures implemented in C, linked to R software. Communication and evaluation of probabilistic, point and interval predictions is investigated. A non-randomized PIT histogram is proposed for correctly assessing calibration and coverage of two-part semicontinuous models. Predictions obtained with the different model specifications are evaluated via graphical tools (Reliability Plot, Sharpness Histogram, PIT Histogram, Brier Score Plot and Quantile Decomposition Plot), proper scoring rules (Brier Score, Continuous Rank Probability Score) and consistent scoring functions (Root Mean Square Error and Mean Absolute Error addressing the predictive mean and median, respectively). Calibration is reached and the inclusion of neighbouring information slightly improves predictions. All specifications outperform a benchmark model with incorrelated effects, confirming the relevance of spatial correlation for modeling rainfall probability and accumulation.

Stereoselective virtual screening of the ZINC database using atom pair 3D-fingerprints

Background Tools to explore large compound databases in search for analogs of query molecules provide a strategically important support in drug discovery to help identify available analogs of any given reference or hit compound by ligand based virtual screening (LBVS). We recently showed that large databases can be formatted for very fast searching with various 2D-fingerprints using the city-block distance as similarity measure, in particular a 2D-atom pair fingerprint (APfp) and the related category extended atom pair fingerprint (Xfp) which efficiently encode molecular shape and pharmacophores, but do not perceive stereochemistry. Here we investigated related 3D-atom pair fingerprints to enable rapid stereoselective searches in the ZINC database (23.2 million 3D structures). Results Molecular fingerprints counting atom pairs at increasing through-space distance intervals were designed using either all atoms (16-bit 3DAPfp) or different atom categories (80-bit 3DXfp). These 3D-fingerprints retrieved molecular shape and pharmacophore analogs (defined by OpenEye ROCS scoring functions) of 110,000 compounds from the Cambridge Structural Database with equal or better accuracy than the 2D-fingerprints APfp and Xfp, and showed comparable performance in recovering actives from decoys in the DUD database. LBVS by 3DXfp or 3DAPfp similarity was stereoselective and gave very different analogs when starting from different diastereomers of the same chiral drug. Results were also different from LBVS with the parent 2D-fingerprints Xfp or APfp. 3D- and 2D-fingerprints also gave very different results in LBVS of folded molecules where through-space distances between atom pairs are much shorter than topological distances. Conclusions 3DAPfp and 3DXfp are suitable for stereoselective searches for shape and pharmacophore analogs of query molecules in large databases. Web-browsers for searching ZINC by 3DAPfp and 3DXfp similarity are accessible at www.gdb.unibe.ch webcite and should provide useful assistance to drug discovery projects.

Identification of protein oligomerization states by analysis of interface conservation

The discrimination of true oligomeric protein–protein contacts from nonspecific crystal contacts remains problematic. Criteria that have been used previously base the assignment of oligomeric state on consideration of the area of the interface and/or the results of scoring functions based on statistical potentials. Both techniques have a high success rate but fail in more than 10% of cases. More importantly, the oligomeric states of several proteins are incorrectly assigned by both methods. Here we test the hypothesis that true oligomeric contacts should be identifiable on the basis of an increased degree of conservation of the residues involved in the interface. By quantifying the degree of conservation of the interface and comparing it with that of the remainder of the protein surface, we develop a new criterion that provides a highly effective complement to existing methods.

Higher order elicitability and Osband’s principle

A statistical functional, such as the mean or the median, is called elicitable if there is a scoring function or loss function such that the correct forecast of the functional is the unique minimizer of the expected score. Such scoring functions are called strictly consistent for the functional. The elicitability of a functional opens the possibility to compare competing forecasts and to rank them in terms of their realized scores. In this paper, we explore the notion of elicitability for multi-dimensional functionals and give both necessary and sufficient conditions for strictly consistent scoring functions. We cover the case of functionals with elicitable components, but we also show that one-dimensional functionals that are not elicitable can be a component of a higher order elicitable functional. In the case of the variance, this is a known result. However, an important result of this paper is that spectral risk measures with a spectral measure with finite support are jointly elicitable if one adds the “correct” quantiles. A direct consequence of applied interest is that the pair (Value at Risk, Expected Shortfall) is jointly elicitable under mild conditions that are usually fulfilled in risk management applications.

MHC class II binding prediction by molecular docking

Proteins of the Major Histocompatibility Complex (MHC) bind self and nonself peptide antigens or epitopes within the cell and present them at the cell surface for recognition by T cells. All T-cell epitopes are MHC binders but not all MCH binders are T-cell epitopes. The MHC class II proteins are extremely polymorphic. Polymorphic residues cluster in the peptide-binding region and largely determine the MHC's peptide selectivity. The peptide binding site on MHC class II proteins consist of five binding pockets. Using molecular docking, we have modelled the interactions between peptide and MHC class II proteins from locus DRB1. A combinatorial peptide library was generated by mutation of residues at peptide positions which correspond to binding pockets (so called anchor positions). The binding affinities were assessed using different scoring functions. The normalized scoring functions for each amino acid at each anchor position were used to construct quantitative matrices (QM) for MHC class II binding prediction. Models were validated by external test sets comprising 4540 known binders. Eighty percent of the known binders are identified in the best predicted 15% of all overlapping peptides, originating from one protein. © 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

Top-K Query Processing in Edge-Labeled Graph Data

Edge-labeled graphs have proliferated rapidly over the last decade due to the increased popularity of social networks and the Semantic Web. In social networks, relationships between people are represented by edges and each edge is labeled with a semantic annotation. Hence, a huge single graph can express many different relationships between entities. The Semantic Web represents each single fragment of knowledge as a triple (subject, predicate, object), which is conceptually identical to an edge from subject to object labeled with predicates. A set of triples constitutes an edge-labeled graph on which knowledge inference is performed. Subgraph matching has been extensively used as a query language for patterns in the context of edge-labeled graphs. For example, in social networks, users can specify a subgraph matching query to find all people that have certain neighborhood relationships. Heavily used fragments of the SPARQL query language for the Semantic Web and graph queries of other graph DBMS can also be viewed as subgraph matching over large graphs. Though subgraph matching has been extensively studied as a query paradigm in the Semantic Web and in social networks, a user can get a large number of answers in response to a query. These answers can be shown to the user in accordance with an importance ranking. In this thesis proposal, we present four different scoring models along with scalable algorithms to find the top-k answers via a suite of intelligent pruning techniques. The suggested models consist of a practically important subset of the SPARQL query language augmented with some additional useful features. The first model called Substitution Importance Query (SIQ) identifies the top-k answers whose scores are calculated from matched vertices' properties in each answer in accordance with a user-specified notion of importance. The second model called Vertex Importance Query (VIQ) identifies important vertices in accordance with a user-defined scoring method that builds on top of various subgraphs articulated by the user. Approximate Importance Query (AIQ), our third model, allows partial and inexact matchings and returns top-k of them with a user-specified approximation terms and scoring functions. In the fourth model called Probabilistic Importance Query (PIQ), a query consists of several sub-blocks: one mandatory block that must be mapped and other blocks that can be opportunistically mapped. The probability is calculated from various aspects of answers such as the number of mapped blocks, vertices' properties in each block and so on and the most top-k probable answers are returned. An important distinguishing feature of our work is that we allow the user a huge amount of freedom in specifying: (i) what pattern and approximation he considers important, (ii) how to score answers - irrespective of whether they are vertices or substitution, and (iii) how to combine and aggregate scores generated by multiple patterns and/or multiple substitutions. Because so much power is given to the user, indexing is more challenging than in situations where additional restrictions are imposed on the queries the user can ask. The proposed algorithms for the first model can also be used for answering SPARQL queries with ORDER BY and LIMIT, and the method for the second model also works for SPARQL queries with GROUP BY, ORDER BY and LIMIT. We test our algorithms on multiple real-world graph databases, showing that our algorithms are far more efficient than popular triple stores.

Improving drug discovery using hybrid softcomputing methods

Virtual screening (VS) methods can considerably aid clinical research, predicting how ligands interact with drug targets. Most VS methods suppose a unique binding site for the target, but it has been demonstrated that diverse ligands interact with unrelated parts of the target and many VS methods do not take into account this relevant fact. This problem is circumvented by a novel VS methodology named BINDSURF that scans the whole protein surface in order to find new hotspots, where ligands might potentially interact with, and which is implemented in last generation massively parallel GPU hardware, allowing fast processing of large ligand databases. BINDSURF can thus be used in drug discovery, drug design, drug repurposing and therefore helps considerably in clinical research. However, the accuracy of most VS methods and concretely BINDSURF is constrained by limitations in the scoring function that describes biomolecular interactions, and even nowadays these uncertainties are not completely understood. In order to improve accuracy of the scoring functions used in BINDSURF we propose a hybrid novel approach where neural networks (NNET) and support vector machines (SVM) methods are trained with databases of known active (drugs) and inactive compounds, being this information exploited afterwards to improve BINDSURF VS predictions.

Improvement of Virtual Screening Predictions using Computational Intelligence Methods

Virtual Screening (VS) methods can considerably aid clinical research, predicting how ligands interact with drug targets. However, the accuracy of most VS methods is constrained by limitations in the scoring function that describes biomolecular interactions, and even nowadays these uncertainties are not completely understood. In order to improve accuracy of scoring functions used in most VS methods we propose a hybrid novel approach where neural networks (NNET) and support vector machines (SVM) methods are trained with databases of known active (drugs) and inactive compounds, this information being exploited afterwards to improve VS predictions.

Descubrimiento de fármacos basado en cribado virtual refinado con enfoques neuronales paralelos

En el campo de la medicina clínica es crucial poder determinar la seguridad y la eficacia de los fármacos actuales y además acelerar el descubrimiento de nuevos compuestos activos. Para ello se llevan a cabo ensayos de laboratorio, que son métodos muy costosos y que requieren mucho tiempo. Sin embargo, la bioinformática puede facilitar enormemente la investigación clínica para los fines mencionados, ya que proporciona la predicción de la toxicidad de los fármacos y su actividad en enfermedades nuevas, así como la evolución de los compuestos activos descubiertos en ensayos clínicos. Esto se puede lograr gracias a la disponibilidad de herramientas de bioinformática y métodos de cribado virtual por ordenador (CV) que permitan probar todas las hipótesis necesarias antes de realizar los ensayos clínicos, tales como el docking estructural, mediante el programa BINDSURF. Sin embargo, la precisión de la mayoría de los métodos de CV se ve muy restringida a causa de las limitaciones presentes en las funciones de afinidad o scoring que describen las interacciones biomoleculares, e incluso hoy en día estas incertidumbres no se conocen completamente. En este trabajo abordamos este problema, proponiendo un nuevo enfoque en el que las redes neuronales se entrenan con información relativa a bases de datos de compuestos conocidos (proteínas diana y fármacos), y se aprovecha después el método para incrementar la precisión de las predicciones de afinidad del método de CV BINDSURF.