Karen T. Schomburg

From structure diagrams to visual chemical patterns.

The intuitive way of chemists to communicate molecules is via two-dimensional structure diagrams. The straightforward visual representations are mostly preferred to the often complicated systematic chemical names. For chemical patterns, however, no comparable visualization standards have evolved so far. Chemical patterns denoting descriptions of chemical features are needed whenever a set of molecules is filtered for certain properties. The currently available representations are constrained to linear molecular pattern languages which are hardly human readable and therefore keep chemists without computational background from systematically formulating patterns. Therefore, we introduce a new visualization concept for chemical patterns. The common standard concept of structure diagrams is extended to account for property descriptions and logic combinations of chemical features in patterns. As a first application of the new concept, we developed the SMARTSviewer, a tool that converts chemical patterns encoded in SMARTS strings to a visual representation. The graphic pattern depiction provides an overview of the specified chemical features, variations, and similarities without needing to decode the often cryptic linear expressions. Taking recent chemical publications from various fields, we demonstrate the wide application range of a graphical chemical pattern language.

Facing the challenges of structure-based target prediction by inverse virtual screening.

Computational target prediction for bioactive compounds is a promising field in assessing off-target effects. Structure-based methods not only predict off-targets, but, simultaneously, binding modes, which are essential for understanding the mode of action and rationally designing selective compounds. Here, we highlight the current open challenges of computational target prediction methods based on protein structures and show why inverse screening rather than sequential pairwise protein-ligand docking methods are needed. A new inverse screening method based on triangle descriptors is introduced: iRAISE (inverse Rapid Index-based Screening Engine). A Scoring Cascade considering the reference ligand as well as the ligand and active site coverage is applied to overcome interprotein scoring noise of common protein-ligand scoring functions. Furthermore, a statistical evaluation of a score cutoff for each individual protein pocket is used. The ranking and binding mode prediction capabilities are evaluated on different datasets and compared to inverse docking and pharmacophore-based methods. On the Astex Diverse Set, iRAISE ranks more than 35% of the targets to the first position and predicts more than 80% of the binding modes with a root-mean-square deviation (RMSD) accuracy of <2.0 Å. With a median computing time of 5 s per protein, large amounts of protein structures can be screened rapidly. On a test set with 7915 protein structures and 117 query ligands, iRAISE predicts the first true positive in a ranked list among the top eight ranks (median), i.e., among 0.28% of the targets.

Fast Protein Binding Site Comparison via an Index-Based Screening Technology

We present TrixP, a new index-based method for fast protein binding site comparison and function prediction. TrixP determines binding site similarities based on the comparison of descriptors that encode pharmacophoric and spatial features. Therefore, it adopts the efficient core components of TrixX, a structure-based virtual screening technology for large compound libraries. TrixP expands this technology by new components in order to allow a screening of protein libraries. TrixP accounts for the inherent flexibility of proteins employing a partial shape matching routine. After the identification of structures with matching pharmacophoric features and geometric shape, TrixP superimposes the binding sites and, finally, assesses their similarity according to the fit of pharmacophoric properties. TrixP is able to find analogies between closely and distantly related binding sites. Recovery rates of 81.8% for similar binding site pairs, assisted by rejecting rates of 99.5% for dissimilar pairs on a test data set containing 1331 pairs, confirm this ability. TrixP exclusively identifies members of the same protein family on top ranking positions out of a library consisting of 9802 binding sites. Furthermore, 30 predicted kinase binding sites can almost perfectly be classified into their known subfamilies.

Benchmark Data Sets for Structure-Based Computational Target Prediction

Structure-based computational target prediction methods identify potential targets for a bioactive compound. Methods based on protein-ligand docking so far face many challenges, where the greatest probably is the ranking of true targets in a large data set of protein structures. Currently, no standard data sets for evaluation exist, rendering comparison and demonstration of improvements of methods cumbersome. Therefore, we propose two data sets and evaluation strategies for a meaningful evaluation of new target prediction methods, i.e., a small data set consisting of three target classes for detailed proof-of-concept and selectivity studies and a large data set consisting of 7992 protein structures and 72 drug-like ligands allowing statistical evaluation with performance metrics on a drug-like chemical space. Both data sets are built from openly available resources, and any information needed to perform the described experiments is reported. We describe the composition of the data sets, the setup of screening experiments, and the evaluation strategy. Performance metrics capable to measure the early recognition of enrichments like AUC, BEDROC, and NSLR are proposed. We apply a sequence-based target prediction method to the large data set to analyze its content of nontrivial evaluation cases. The proposed data sets are used for method evaluation of our new inverse screening method iRAISE. The small data set reveals the methods capability and limitations to selectively distinguish between rather similar protein structures. The large data set simulates real target identification scenarios. iRAISE achieves in 55% excellent or good enrichment a median AUC of 0.67 and RMSDs below 2.0 Å for 74% and was able to predict the first true target in 59 out of 72 cases in the top 2% of the protein data set of about 8000 structures.

Interactive design of generic chemical patterns

Every medicinal chemist has to create chemical patterns occasionally for querying databases, applying filters or describing functional groups. However, the representations of chemical patterns have been so far limited to languages with highly complex syntax, handicapping the application of patterns. Graphic pattern editors similar to chemical editors can facilitate the work with patterns. In this article, we review the interfaces of frequently used web search engines for chemical patterns. We take a look at pattern editing concepts of standalone chemical editors and finally present a completely new, unpublished graphical approach to pattern design, the SMARTSeditor.

Computational biotechnology: prediction of competitive substrate inhibition of enzymes by buffer compounds with protein-ligand docking.

In vitro enzymatic activity highly depends on the reaction medium. One of the most important parameters is the buffer used to keep the pH stable. The buffering compound prevents a severe pH-change and therefore a possible denaturation of the enzyme. However buffer agents can also have negative effects on the enzymatic activity, such as competitive substrate inhibition. We assess this effect with a computational approach based on a protein-ligand docking method and the HYDE scoring function. Our method predicts competitive binding of the buffer compound to the active site of the enzyme. Using data from literature and new experimental data, the procedure is evaluated on nine different enzymatic reactions. The method predicts buffer-enzyme interactions and is able to score these interactions with the correct trend of enzymatic activities. Using the new method, possible buffers can be selected or discarded prior to laboratory experiments.

Large-Scale Analysis of Hydrogen Bond Interaction Patterns in Protein–Ligand Interfaces

Protein-ligand interactions are the fundamental basis for molecular design in pharmaceutical research, biocatalysis, and agrochemical development. Especially hydrogen bonds are known to have special geometric requirements and therefore deserve a detailed analysis. In modeling approaches a more general description of hydrogen bond geometries, using distance and directionality, is applied. A first study of their geometries was performed based on 15 protein structures in 1982. Currently there are about 95 000 protein-ligand structures available in the PDB, providing a solid foundation for a new large-scale statistical analysis. Here, we report a comprehensive investigation of geometric and functional properties of hydrogen bonds. Out of 22 defined functional groups, eight are fully in accordance with theoretical predictions while 14 show variations from expected values. On the basis of these results, we derived interaction geometries to improve current computational models. It is expected that these observations will be useful in designing new chemical structures for biological applications.

Discriminative Chemical Patterns: Automatic and Interactive Design

The classification of molecules with respect to their inhibiting, activating, or toxicological potential constitutes a central aspect in the field of cheminformatics. Often, a discriminative feature is needed to distinguish two different molecule sets. Besides physicochemical properties, substructures and chemical patterns belong to the descriptors most frequently applied for this purpose. As a commonly used example of this descriptor class, SMARTS strings represent a powerful concept for the representation and processing of abstract chemical patterns. While their usage facilitates a convenient way to apply previously derived classification rules on new molecule sets, the manual generation of useful SMARTS patterns remains a complex and time-consuming process. Here, we introduce SMARTSminer, a new algorithm for the automatic derivation of discriminative SMARTS patterns from preclassified molecule sets. Based on a specially adapted subgraph mining algorithm, SMARTSminer identifies structural features that are frequent in only one of the given molecule classes. In comparison to elemental substructures, it also supports the consideration of general and specific SMARTS features. Furthermore, SMARTSminer is integrated into an interactive pattern editor named SMARTSeditor. This allows for an intuitive visualization on the basis of the SMARTSviewer concept as well as interactive adaption and further improvement of the generated patterns. Additionally, a new molecular matching feature provides an immediate feedback on a patterns matching behavior across the molecule sets. We demonstrate the utility of the SMARTSminer functionality and its integration into the SMARTSeditor software in several different classification scenarios.

Facing the challenges of computational target prediction

To which proteins does a compound bind? Is it selective or promiscuous? These are questions which can be answered by inverse virtual screening, which identifies potential targets for a molecule of interest. Experimental screening of a molecule on thousands of targets is costly and elaborate. Contrarily, structure-based computational methods are only limited by the availability of 3D structures, rendering them an important complement to experimental methods. Our inverse screening method XxirT combines triangle descriptor matching [1] with a new ranking approach, considering a reference score for each pocket. A precalculated bitmap encoding of the descriptors and an efficient design of a database for 3D protein structures allows a rapid screening of thousands of protein-ligand complexes with a query compound. Classical protein-ligand scoring functions are not capable of inter-target score comparison, since the absolute values are target-dependent. Therefore, in inverse virtual screening, a major problem is the design of a ranking scheme allowing the comparison of target scores with respect to one query compound [2]. A lack of available data for statistical evaluation of the ranking capability further complicates the task. Data sets mostly contain positive hits, e.g. binding affinities for one molecule to several proteins while lacking negative data points, such as ‘the molecule does not bind to this protein’. As a basis for statistical evaluation of our new inverse screening concept, we introduce a dataset consisting of a ligand set of approved drugs and the scPDB target database [3]. Drugs are well qualified for use in a method evaluation dataset, as they are well tested and had to pass selectivity standard tests to get approved. Therefore, their targets are comparably well-characterized, allowing a classification into true positives and true negatives. This approach is the first that evaluates a structure-based inverse screening method on a systematic statistical test.

Chemical pattern visualization in 2D – the SMARTSviewer

Chemical patterns are essential for various fields of chemical, chemoinformatic and pharmaceutical applications. So far, representations of chemical patterns are limited to linear molecular pattern languages like SMARTS[1]. As these languages are designed for computational efficiency, they are often hardly human readable. In order to improve the usability of chemical patterns for scientists without expert knowledge of one of these languages, we present a visual representation of chemical patterns similar to structure diagrams. While molecules can also be represented by systematic names, the means of communication of compounds among scientists is the visual representation of 2D structure diagrams. Therefore, we propose a depiction of chemical patterns based on the common standard concept of structure diagrams. As chemical patterns denote descriptions of chemical features, the concept of structure diagrams is extended with graphical elements to depict property descriptions and logic combinations of chemical features. The aim of the depiction is to provide an overview of the specified features as well as to highlight similarities and differences among patterns. As a first application of the new visualization concept we developed the SMARTSviewer. The tool converts a pattern in form of a SMARTS string into a graphic representation. Along with the graphic depiction, the tool produces a legend explaining the graphic symbols and meaning of the features described in the pattern. The SMARTSviewer is openly accessible [2,3]. Since commonly accepted visual depictions have to evolve from the needs of the users, we hope to initiate a discussion based on the concept we introduce.