Katrin Stierand

Drawing the PDB: Protein-Ligand Complexes in Two Dimensions.

The two-dimensional representation of molecules is a popular communication medium in chemistry and the associated scientific fields. Computational methods for drawing small molecules with and without manual investigation are well-established and widely spread in terms of numerous software tools. Concerning the planar depiction of molecular complexes, there is considerably less choice. We developed the software PoseView, which automatically generates two-dimensional diagrams of macromolecular complexes, showing the ligand, the interactions, and the interacting residues. All depicted molecules are drawn on an atomic level as structure diagrams; thus, the output plots are clearly structured and easily readable for the scientist. We tested the performance of PoseView in a large-scale application on nearly all druglike complexes of the PDB (approximately 200000 complexes); for more than 92% of the complexes considered for drawing, a layout could be computed. In the following, we will present the results of this application study.

Molecular complexes at a glance

MOTIVATION In this paper a new algorithmic approach is presented, which automatically generates structure diagrams of molecular complexes. A complex diagram contains the ligand, the amino acids of the protein interacting with the ligand and the hydrophilic interactions schematized as dashed lines between the corresponding atoms. The algorithm is based on a combinatorial optimization strategy which solves parts of the layout problem non-heuristically. The depicted molecules are represented as structure diagrams according to the chemical nomenclature. Due to the frequent usage of complex diagrams in the scientific literature as well as in text books dealing with structural biology, biochemistry and medicinal chemistry, the new algorithm is a key element for computer applications in these areas. RESULTS The method was implemented in the new software tool PoseView. It was tested on a representative dataset containing 305 protein-ligand complexes in total from the Brookhaven Protein Data Bank. PoseView was able to find collision-free layouts for more than three quarters of all complexes. In the following the layout generation algorithm is presented and, additional to the statistical results, representative test cases demonstrating the challenges of the layout generation will be discussed. AVAILABILITY The method is available as a webservice at http://www.zbh.uni-hamburg.de/poseview.

From Modeling to Medicinal Chemistry: Automatic Generation of Two-Dimensional Complex Diagrams

As a result of the increasing application of structure‐based drug design, the visualization of protein–ligand complexes has become an important feature in medicinal chemistry. The large number of experimentally resolved complex structures and the further development of computer‐aided methods like docking or de novo design establishes new possibilities in this field. During lead finding and optimization, a manual investigation of many complexes and their interaction patterns is typically performed. We present an algorithm that automatically generates 2D‐protein–ligand diagrams as a possible solution for a transparent visualization of the contact partners in a complex and as a support for scientists in the evaluation of structure‐based design results. Running the software on representative test data sets, it generates collision free layouts for ∼76 % of the cases in the range of tenths of a second per complex. The success rate for complexes with ligands which have a molecular weight <500 Da is 87 %.

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.

PoseView -- molecular interaction patterns at a glance

Chemists are well trained in perceiving 2D molecular sketches. On the side of computer assistance, the automated generation of such sketches becomes very difficult when it comes to multi-molecular arrangements such as protein-ligand complexes in a drug design context. Existing solutions to date suffer from drawbacks such as missing important interaction types, inappropriate levels of abstraction and layout quality. During the last few years we have developed PoseView [1,2], a tool which displays molecular complexes incorporating a simple, easy-to-perceive arrangement of the ligand and the amino acids towards which it forms interactions. Resulting in atomic resolution diagrams, PoseView operates on a fast tree re-arrangement algorithm to minimize crossing lines in the sketches. Due to a de-coupling of interaction perception and the drawing engine, PoseView can draw any interactions determined by either distance-based rules or the FlexX interaction model (which itself is user accessible). Owing to the small molecule drawing engine 2Ddraw [3], molecules are drawn in a textbook-like manner following the IUPAC regulations. The tool has a generic file interface for other complexes than protein-ligand arrangements. It can therefore be used as well for the display of, e.g., RNA/DNA complexes with small molecules. For batch processing, an additional command line interface is available; output can be provided in various formats, amongst them gif, ps, svg and pdf. Besides the underlying interaction models, we will present new algorithmic approaches, assess usability issues and a large-scale validation study on the PDB.

The Internet as Scientific Knowledge Base: Navigating the Chem-Bio Space

The early phases of designing a new drug are characterized by an exhaustive exploration of related information available in public as well as in proprietary databases. Gathering such information for a given set of compounds is complicated due to the large number of data sources, the different data formats and query mechanisms used.1 Manual searches are tedious and time consuming, thus usually limited to individual compounds only. The process is error prone and the information collected this way is incomplete, of variable quality, and missing the link to the original data sources. Therefore data values can not necessarily be compared and it is difficult to identify compound duplicates.2 A recent approach to overcome these issues is the integration of data from different sources by means of semantic web technologies.3 Organizing the data in triples allows the representation of relationships between data points. Here, each triple has a subject-predicate-object structure, i.e. “sodium chloride (=subject) is a (=predicate) salt (=object)”. This way, related pieces of information can be connected semantically. Next to assembling the data itself, another challenging aspect is the representation of the data in a user-friendly and comprehensible manner. The constantly growing amount of data to be considered in a single drug discovery process poses new challenges to find a human accessible representation. The visualization of compound structures along with the related textual information should be designed for interactive use and at the same time provide access to important details. Commonly used cheminformatics tools for the handling of large sets of compounds include Spotfire (TIBCO),4 MS Excel,5 Accord (Accelrys),6 the Dotmatics Browser,7 Seurat (Schrodinger),8 CDD Vault,9 and Tableau.10 These tools allow exploring entire datasets, e.g. by plotting numerical properties against each other or by presenting them in a tabular form. However, the compound information cannot be augmented dynamically, but is instead restricted to the information contained in the original dataset. Employing the Internet as a knowledge base entails different challenges. The available data is spread over numerous sources, which are updated frequently but not necessarily regularly. Due to this dynamic nature of data, downloading a copy and working with it locally quickly leads to outdated, false or simply missing information. Other, more technical challenges are the large amounts of data, the diverse and often non-standard formats along with the use of different molecule identifiers in most data sources. In the context of semantic web technology, this can be solved by so-called identity resolution services (IRS),11 which are entrusted with the mapping of different identifiers to one unique object representation. In 2011 a new project, called OpenPHACTS (Open Pharmacological Concept Triple Store),12 has been initiated aiming to create an Open Pharmacological Space (OPS) by assembling data useful for drug discovery from public data sources using semantic web technologies. Currently, it includes datasets from the following databases: DrugBank,13 ChEMBL,14 SwissProt/UniProt,15 ChEBI,16 Gene Ontology,17 GOA,18 Wikipathways.19 Following an application-oriented approach, the project started with the definition of potential use cases in the form of research questions, formulated and prioritized by a consortium of scientists from both academia as well as various pharmaceutical companies.20 The development of the different software components is now led by these research questions. So-called Exemplar Services are an important part of the OpenPHACTS project, show-casing the versatile application possibilities of OPS. These web services are using the OPS framework and provide access to the data in a user friendly manner. Each exemplar focusses on a certain application domain, covering target-related, polypharmacology as well as compound-related research questions.20 The exemplar services should enable the average bench scientist to access the data in an intuitive fashion with minimal learning effort and particularly avoiding the necessity to use complicated query languages such as JavaScript/JSON or SPARQL.21 Here we present a first prototype of the ChemBioNavigator (CBN), an OpenPHACTS exemplar service for navigating the chem-bio space with a focus on small molecules relevant in pharmaceutical research. This service allows to access large amounts of data originating from numerous public data sources available on the Internet and to merge this with proprietary compound information dynamically during runtime. The added information is taken directly from datasets included in the OPS or from external data sources which are referenced in the OPS data cache. Taking a step beyond the use-case driven development, the CBN is based on the analysis of several interviews with potential users from pharmaceutical industry. The interviews gave invaluable insights not only about the day to day use-cases, but also deficiencies as well as highly valued features of the existing tools. While the requirements of users may be competing, thus can never all be satisfied, trends usually become apparent as the number of interviewed parties grows. This leads to an agile and target-oriented development, which is able to quickly adapt to changing requirements from the scientist’s daily work. The CBN is realized using modern web technologies and state of the art cheminformatics software libraries. The user interface consists of two different regions: A large molecule visualization canvas and an information panel (see Figure ​Figure11). Figure 1 The ChemBioNavigator user interface consists of two areas: the information panel on the left hand side and the visualization area on the right hand side. When a compound is selected in the drawing area, it is highlighted by a green circle and the available ... A CBN-session starts by uploading molecules either as a SD file or a mol2 file or in the form of SMILES strings. These are processed and validated using the NAOMI22 software library. NAOMI also calculates certain base-properties, such as the molecular weight, the number of hydrogen bond donors and acceptors, or a calculated logP. Subsequently the compounds are annotated with information obtained from the OPS system. These additional information range from identifiers and one-dimensional compound representations, such as InChI23 and SMILES24 strings, over topological and physico-chemical properties, to target-relations and assay data. Compound data uploaded from file is displayed and available for analysis as well. For data analysis purpose, the entire compound set can be drawn in a scatter-plot on the visualization canvas. The user can choose any numerical property from any of the incorporated data sources to sort-order the data points via the x- and y-axis of the scatter plot. The data is color-coded based on whether the system was able to identify the compound within OPS and was able to obtain additional information. As known from browser based map services, it is possible to move through the scatter-plot using the mouse as well as to zoom in and out. When zooming in, the level of detail increases and mere dots are replaced by the actual molecule depictions. Additionally, a tabular representation of structure diagrams and the detailed view of single compounds are available allowing a closer examination of selected molecules. From the set of compounds, subsets can be selected and saved as SD file for further use. The SD file covers all compound data including the information retrieved from the OPS platform. The information panel is subdivided into the following tabs: The details-tab shows the structure diagram of the currently selected compound and lists all available properties. The properties are grouped as base properties assigned during the initialization process, properties provided in the uploaded file and additional properties originating from the OPS platform. The selection of properties for the axes of the scatterplot is visually supported by histograms, generated on the fly showing the value range and distribution of the currently chosen property. Using the OPS system as data source, the CBN provides homogenous access to heterogeneous data originating from various databases. The functionality described here constitutes a first prototype of the CBN, which will be extended with various additional features over the course of the OpenPHACTS project. The main focus of further development will be the inclusion of target- and biological data. Given the project-focus on pharmaceutical research, the relationship between compound and potential targets is of particular interest. However, rather than duplicating functionality, links into other exemplars shall be provided to address very detailed or more complex research questions from other domains. Another important feature for future releases is the possibility to extend a compound set based on similarity searches. The incorporation of these and other new features will enable the average bench scientist to easily explore large amounts of data, augmenting their research results with new aspects. The CBN is made publicly available at http://cbn.zbh.uni-hamburg.de.

Consistent two-dimensional visualization of protein-ligand complex series

BackgroundThe comparative two-dimensional graphical representation of protein-ligand complex series featuring different ligands bound to the same active site offers a quick insight in their binding mode differences. In comparison to arbitrary orientations of the residue molecules in the individual complex depictions a consistent placement improves the legibility and comparability within the series. The automatic generation of such consistent layouts offers the possibility to apply it to large data sets originating from computer-aided drug design methods.ResultsWe developed a new approach, which automatically generates a consistent layout of interacting residues for a given series of complexes. Based on the structural three-dimensional input information, a global two-dimensional layout for all residues of the complex ensemble is computed. The algorithm incorporates the three-dimensional adjacencies of the active site residues in order to find an universally valid circular arrangement of the residues around the ligand. Subsequent to a two-dimensional ligand superimposition step, a global placement for each residue is derived from the set of already placed ligands. The method generates high-quality layouts, showing mostly overlap-free solutions with molecules which are displayed as structure diagrams providing interaction information in atomic detail. Application examples document an improved legibility compared to series of diagrams whose layouts are calculated independently from each other.ConclusionsThe presented method extends the field of complex series visualizations. A series of molecules binding to the same protein active site is drawn in a graphically consistent way. Compared to existing approaches these drawings substantially simplify the visual analysis of large compound series.

Flat and Easy: 2D Depiction of Protein-Ligand Complexes.

Visualization of molecular complexes is commonly used to support the investigation of interaction patterns formed between the members of a molecular complex ensemble. Similar to the representation of single small molecules as structure diagrams, a schematic two‐dimensional design of molecular complexes features several advantages. This visualization mode enables the peer to scan large numbers of complexes in short time, originating for example from a virtual screening or de novo design campaign, and to get an impression of their quality. In addition, the diagrams can be printed on hardcopies without information loss and are therefore well suited for publications and talks. We will give an overview of the existing algorithms for the automatic generation of two‐dimensional complex diagrams. The reduction of dimensions from three to two is a quite difficult task since the resulting layout has to be more or less free of overlaps and has to follow esthetical guidelines. All programs proceed on input being composed of the three‐dimensional ligand and receptor coordinates. Due to the lack of guiding principles, the algorithms and resulting diagrams of the different available tools substantially differ in graphical styles, level of detail, and information content.

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.

Accessing Open PHACTS: interactive exploration of compounds and targets from the semantic web

Pharmacological research is hampered by scattered data which have to be retrieved by varying methods and in different data formats. This heterogeneity increases research costs and limits throughput. Over the last two years, the Open PHACTS Discovery Platform [1] has been developed as a centralized repository, integrating pharmacological data from a variety of information resources and providing tools and services to query these integrated data in pharmacological research. Following an application-oriented approach, the Open PHACTS project started with the definition of potential use cases in the form of prioritized research questions [2], most of which can only be answered by accessing multiple data sources in the web. The development of the platform as well as the services has been guided by these questions. Here, we present the ChemBioNavigator (CBN) [3], a web application allowing to navigate the Open PHACTS chem-bio space with a focus on small molecules and their targets. CBN comprises of a large visualization area with different view modes and two information panels, allowing a deeper insight in information for compounds and targets. It allows interactive exploration of compound sets through sorting and subset selection as well as extending sets by substructure or similarity search. The relation between compounds and targets is defined by assay data from the Discovery Platform. Each compound and each target is annotated with information from multiple data sources which is provided together with the provenance for each data point. In this contribution we roughly outline the OpenPHACTS/CBN technology and present a number of high-priority research questions, highlight the advantages of exploiting the integrated data through the CBNs smart and intuitive interface.