- Molecular similarity for machine learning in drug development : poster presentation (2008)
- Poster presentation In pharmaceutical research and drug development, machine learning methods play an important role in virtual screening and ADME/Tox prediction. For the application of such methods, a formal measure of similarity between molecules is essential. Such a measure, in turn, depends on the underlying molecular representation. Input samples have traditionally been modeled as vectors. Consequently, molecules are represented to machine learning algorithms in a vectorized form using molecular descriptors. While this approach is straightforward, it has its shortcomings. Amongst others, the interpretation of the learned model can be difficult, e.g. when using fingerprints or hashing. Structured representations of the input constitute an alternative to vector based representations, a trend in machine learning over the last years. For molecules, there is a rich choice of such representations. Popular examples include the molecular graph, molecular shape and the electrostatic field. We have developed a molecular similarity measure defined directly on the (annotated) molecular graph, a long-standing established topological model for molecules. It is based on the concepts of optimal atom assignments and iterative graph similarity. In the latter, two atoms are considered similar if their neighbors are similar. This recursive definition leads to a non-linear system of equations. We show how to iteratively solve these equations and give bounds on the computational complexity of the procedure. Advantages of our similarity measure include interpretability (atoms of two molecules are assigned to each other, each pair with a score expressing local similarity; this can be visualized to show similar regions of two molecules and the degree of their similarity) and the possibility to introduce knowledge about the target where available. We retrospectively tested our similarity measure using support vector machines for virtual screening on several pharmaceutical and toxicological datasets, with encouraging results. Prospective studies are under way.
- Kernel learning for ligand-based virtual screening:discovery of a new PPARgamma agonist (2010)
- Poster presentation at 5th German Conference on Cheminformatics: 23. CIC-Workshop Goslar, Germany. 8-10 November 2009 We demonstrate the theoretical and practical application of modern kernel-based machine learning methods to ligand-based virtual screening by successful prospective screening for novel agonists of the peroxisome proliferator-activated receptor gamma (PPARgamma) . PPARgamma is a nuclear receptor involved in lipid and glucose metabolism, and related to type-2 diabetes and dyslipidemia. Applied methods included a graph kernel designed for molecular similarity analysis , kernel principle component analysis , multiple kernel learning , and, Gaussian process regression . In the machine learning approach to ligand-based virtual screening, one uses the similarity principle  to identify potentially active compounds based on their similarity to known reference ligands. Kernel-based machine learning  uses the "kernel trick", a systematic approach to the derivation of non-linear versions of linear algorithms like separating hyperplanes and regression. Prerequisites for kernel learning are similarity measures with the mathematical property of positive semidefiniteness (kernels). The iterative similarity optimal assignment graph kernel (ISOAK)  is defined directly on the annotated structure graph, and was designed specifically for the comparison of small molecules. In our virtual screening study, its use improved results, e.g., in principle component analysis-based visualization and Gaussian process regression. Following a thorough retrospective validation using a data set of 176 published PPARgamma agonists , we screened a vendor library for novel agonists. Subsequent testing of 15 compounds in a cell-based transactivation assay  yielded four active compounds. The most interesting hit, a natural product derivative with cyclobutane scaffold, is a full selective PPARgamma agonist (EC50 = 10 ± 0.2 microM, inactive on PPARalpha and PPARbeta/delta at 10 microM). We demonstrate how the interplay of several modern kernel-based machine learning approaches can successfully improve ligand-based virtual screening results.
- Virtuelles Screening und De-Novo-Design von PPARalpha-Agonisten mit Oberflächen-Deskriptoren (2008)
- Die Komplementarität der molekularen Oberflächen und der Pharmakophorpunkte ist ein verbreiteter Konzept im rechnergestützen Moleküldesign. Diesem Konzept folgend wurde die Software SQUIRREL neu entwickelt und in der Programmiersprache Java implemetiert. Die Software generiert die Vorschläge für den bioisosteren Ersatz von Molekülen und Molekülfragmenten. SQUIRREL kombiniert Oberflächen- und Pharmakophoreigenschaften bioaktiver Substanzen und kann im virtuellen Screening und fragment-basierten de novo Design eingesetzt werden. In einer prospektiven Studie wurde SQUIRREL verwendet, um neue selektive PPARalpha-Agonisten aus einer kommerziellen Moleküldatenbank zu identifizieren. Die Software lieferte eine potente Substanz (EC50 = 44 nM) mit über 100facher Selektivität gegenüber PPARgamma. In einer zweiten Studie wurde eine Leitstruktur de novo generiert und synthetisiert. Als Ausgangstruktur diente der bekannte PPARalpha-Agonist GW590735. Während des Designvorgangs wurden zwei Teilstrukturen, die für die Aktivität von GW590735 verantwortlich sind, durch bioisostere Gruppen ersetzt, die von SQUIRRELnovo vorgeschlagen wurden. Die neue Leitstruktur aktiviert PPARalpha in einem zellbasierten Reportergen-Testsystem bei einem EC50 von 0.51 µM.
- SQUIRRELnovo : de novo design of a PPARalpha agonist by bioisosteric replacement (2009)
- Shape complementarity is a compulsory condition for molecular recognition . In our 3D ligand-based virtual screening approach called SQUIRREL, we combine shape-based rigid body alignment  with fuzzy pharmacophore scoring . Retrospective validation studies demonstrate the superiority of methods which combine both shape and pharmacophore information on the family of peroxisome proliferator-activated receptors (PPARs). We demonstrate the real-life applicability of SQUIRREL by a prospective virtual screening study, where a potent PPARalpha agonist with an EC50 of 44 nM and 100-fold selectivity against PPARgamma has been identified. SQUIRREL molecular superposition is based on a graph-matching routine  and allows partial matching. We used this advantage for searching for bioisosteric replacement suggestions in a database of molecular fragments derived from a collection of drug-like compounds . The bioisosteric groups suggested by our tool SQURRELnovo, can be used for ligand-based de novo design by a human expert. Using the fibrate derivative GW590735  as query, we designed a novel lead structure by substitution of the acidic head group and hydrophobic tail. The synthesis and following testing in a cell-based reporter gene assay [7,8] revealed that the designed structure activates PPARalpha with an EC50 of 510 nM.
- PhAST : pharmacophore alignment search tool (2009)
- We developed the Pharmacophore Alignment Search Tool (PhAST), a text-based technique for rapid hit and lead structure searching in large compound databases. For each molecule, a two-dimensional graph of potential pharmacophoric points (PPPs) is created, which has an identical topology as the original molecule with implicit hydrogen atoms. Each vertex is coloured by a symbol representing the corresponding PPP. The vertices of the graph are canonically labelled . The symbols associated with the vertices are combined to a so-called PhAST-Sequence beginning with the vertex with the lowest canonical label. Due to the canonical labelling the created PhAST-Sequence is characteristic for each molecule. For similarity assessment, PhAST-Sequences are compared using the sequence identity in their global pairwise alignment . The alignment score lies between 0 (no similarity) and 1 (identical PhAST-Sequences). In order to use global pairwise sequence alignment, a score matrix for pharmacophoric symbols was developed and gap penalties were optimized. PhAST performed comparably and sometimes superior to other similarity search tools (CATS2D , MOE pharmacophore quadruples ) in retrospective virtual screenings using the COBRA  collection of drugs and lead structures. Most importantly, the PhAST alignment technique allows for the computation of significance estimates that help prioritize a virtual hit list.
- Pseudoreceptor-based pocket selection in a molecular dynamics simulation of the histamine H4 receptor (2009)
- There is a renewed interest in pseudoreceptor models which enable computational chemists to bridge the gap of ligand- and receptor-based drug design . We developed a pseudoreceptor model for the histamine H4 receptor (H4R) based on five potent antagonists representing different chemotypes. Here we present the selection of potential ligand binding pockets that occur during molecular dynamics (MD) simulations of a homology-based receptor model. We present a method for prioritizing receptor models according to their match with the consensus ligand-binding mode represented by the pseudoreceptor. In this way, ligand information can be transferred to receptor-based modelling. We use Geometric Hashing to match three-dimensional points in Cartesion space . This allows for the rapid translation- and rotation-free comparison of atom coordinates, which also permits partial matching. The only prerequisite is a hash table, which uses distance triplets as hash keys. Each time a distance triplet occurring in the candidate point set which corresponds to an existing key, the match is represented by a vote of the respective key. Finally, the global match of both point sets can be easily extracted by selection of voted distance triplets. The results revealed a preferred ligand-binding pocket in H4R, which would not have been identified using an unrefined homology model of the protein. The key idea was to rely on ligand information by pseudoreceptor modelling.
- Virtual chemical reactions for drug design (2009)
- Two methods for the fast, fragment-based combinatorial molecule assembly were developed. The software COLIBREE® (Combinatorial Library Breeding) generates candidate structures from scratch, based on stochastic optimization . Result structures of a COLIBREE design run are based on a fixed scaffold and variable linkers and side-chains. Linkers representing virtual chemical reactions and side-chain building blocks obtained from pseudo-retrosynthetic dissection of large compound databases are exchanged during optimization. The process of molecule design employs a discrete version of Particle Swarm Optimization (PSO) . Assembled compounds are scored according to their similarity to known reference ligands. Distance to reference molecules is computed in the space of the topological pharmacophore descriptor CATS . In a case study, the approach was applied to the de novo design of potential peroxisome proliferator-activated receptor (PPAR gamma) selective agonists. In a second approach, we developed the formal grammar Reaction-MQL  for the in silico representation and application of chemical reactions. Chemical transformation schemes are defined by functional groups participating in known organic reactions. The substructures are specified by the linear Molecular Query Language (MQL) . The developed software package contains a parser for Reaction-MQL-expressions and enables users to design, test and virtually apply chemical reactions. The program has already been used to create combinatorial libraries for virtual screening studies. It was also applied in fragmentation studies with different sets of retrosynthetic reactions and various compound libraries.
- Identification of Plk1 type II inhibitors by structure-based virtual screening (2009)
- Protein kinases are targets for drug development . Dysregulation of kinase activity leads to various diseases , e.g. cancer, inflammation, diabetes . Human polo-like kinase 1 (Plk1), a serine/threonine kinase, is a cancer-relevant gene and a potential drug target which attracts increasing attention in the field of cancer therapy. Plk1 is a key player in mitosis and modulates entry into mitosis and the spindle checkpoint at the meta-/anaphase transition. Plk1 overexpression is observed in various human tumors, and it is a negative prognostic factor for cancer patients . The same catalytical mechanism and the same co-substrate (ATP) lead to the problem of inhibitor selectivity. A strategy to solve this problem is represented by targeting the inactive conformation of kinases . Kinases undergo conformational changes between active and inactive conformation and thus an additional hydrophobic pocket is created in the inactive conformation where the surrounding amino acids are less conserved . A "homology model" of the inactive conformation of Plk1 was constructed, as the crystal structure in its inactive conformation is unknown. A crystal structure of Aurora A kinase served as template structure. With this homology model a receptor-based pharmacophore search was performed using SYBYL7.3 software. The raw hits were filtered using physico-chemical properties. The resulting hits were docked using Gold3.2 software, and 13 candidates for biological testing were manually selected. Three compounds of the 13 tested exhibit anti-proliferative effects in HeLa cancer cells. The most potent inhibitor, SBE13, was further tested in various other cancer cell lines of different origins and displayed EC50 values between 12 microM and 39 microM. Cancer cells incubated with SBE13 showed induction of apoptosis, detected by PARP (Poly-Adenosyl-Ribose-Polymerase) cleavage, caspase 9 activation and DAPI staining of apoptotic nuclei.
- SBE13, a newly identified inhibitor of inactive polo-like kinase 1 (2010)
- Poster presentation at 5th German Conference on Cheminformatics: 23. CIC-Workshop Goslar, Germany. 8-10 November 2009 Protein kinases are important targets for drug development. The almost identical protein folding of kinases and the common co-substrate ATP leads to the problem of inhibitor selectivity. Type II inhibitors, targeting the inactive conformation of kinases, occupy a hydrophobic pocket with less conserved surrounding amino acids . Human polo-like kinase 1 (Plk1) represents a promising target for approaches to identify new therapeutic agents. Plk1 belongs to a family of highly conserved serine/threonine kinases, and is a key player in mitosis, where it modulates the spindle checkpoint at metaphase/anaphase transition. Plk1 is over-expressed in all today analyzed human tumors of different origin and serves as a negative prognostic marker in cancer patients. The newly identified inhibitor, SBE13, a vanillin derivative, targets Plk1 in its inactive conformation . This leads to selectivity within the Plk family and towards Aurora A. This selectivity can be explained by docking studies of SBE13 into the binding pocket of homology models of Plk1, Plk2 and Plk3 in their inactive conformation. SBE13 showed anti-proliferative effects in cancer cell lines of different origins with EC50 values between 5 microM and 39 microM and induced apoptosis. Increasing concentrations of SBE13 result in increasing amounts of cells in G2/M phase 13 hours after double thymidin block of HeLa cells. The kinase activity of Plk1 was inhibited with an IC50 of 200 pM. Taken together, we could show that carefully designed structure-based virtual screening is well-suited to identify selective type II kinase inhibitors targeting Plk1 as potential anti-cancer therapeutics.
- DOGS: reaction-driven de novo design of bioactive compounds (2012)
- We present a computational method for the reaction-based de novo design of drug-like molecules. The software DOGS (Design of Genuine Structures) features a ligand-based strategy for automated ‘in silico’ assembly of potentially novel bioactive compounds. The quality of the designed compounds is assessed by a graph kernel method measuring their similarity to known bioactive reference ligands in terms of structural and pharmacophoric features. We implemented a deterministic compound construction procedure that explicitly considers compound synthesizability, based on a compilation of 25'144 readily available synthetic building blocks and 58 established reaction principles. This enables the software to suggest a synthesis route for each designed compound. Two prospective case studies are presented together with details on the algorithm and its implementation. De novo designed ligand candidates for the human histamine H4 receptor and γ-secretase were synthesized as suggested by the software. The computational approach proved to be suitable for scaffold-hopping from known ligands to novel chemotypes, and for generating bioactive molecules with drug-like properties.