SeeSAR/HYDE: scientific background

SeeSAR has not only the most MedChem-user friendly graphical interface but is also loaded with the most cutting edge science and technology. You will have heard about the HYDE affinity assessment, ReCore for re-scaffolding, and FlexX-docking but there is more. Thanks to a tremendous collaboration with our outstanding academic partners we can serve our users with the best in class technology. Read on for more details and references to the primary literature.


google-like re-scaffolding

Task: Generate new IP or get rid of a problem with a molecule, specify bonds or interactions to be matched by new fragments. The arrangement of these vectors is taken to a fragment library that has been pre-processed for speed ("indexed"). From this, results are retrieved within fractions of a second, and using a 4-dimensional vector, the quality of the fit is computed. The fragment library ("index") can be generated from crystals or computer-generated files, or the user can certainly also index corporate structures. Fragment growing with one vector and fragment merging/linking is supported in LeadIT. Even small indices capture millions of possibilities by design. Fast, local SDDs will boost the speed of delivery.
ReCore emerged from a collaboration with Roche Basle and Hamburg University and has been much augmented and extended by BioSolveIT thereafter.

Recore: A Fast and Versatile Method for Scaffold Hopping Based on Small Molecule Crystal Structure Conformations
Patrick Maass, Tanja Schulz-Gasch, Martin Stahl, and Matthias Rarey
Journal of Chemical Information and Modeling 2007 47 (2), 390-399
ReCore (Computer Software Review)
John H. VanDrie
Journal of the American Chemical Society 2009 131 (4), 1617-1617

visual torsions

statistical assessment of likelihood of dihedrals

Task: Based on a rigorous statistical analysis of the CCDC small molecules crystal structure database of thousands of molecules, the traffic-light implementation for the torsion angles in molecules reflects the frequency of occurrence of a given dihedral. The underlying assumption is that frequent observations correlate with low energy structures and vice versa (Inverse Boltzmann). The computation is so fast, that during editing, the statistical feedback is given instantaneously.
The Visual Torsions emerged from a collaboration with Roche Basle and Hamburg University.

Torsion Angle Preferences in Druglike Chemical Space: A Comprehensive Guide
Christin Schärfer, Tanja Schulz-Gasch, Hans-Christian Ehrlich, Wolfgang Guba, Matthias Rarey, and Martin Stahl
Journal of Medicinal Chemistry 2013 56 (5), 2016-2028
Torsion Library Reloaded: A New Version of Expert-Derived SMARTS Rules for Assessing Conformations of Small Molecules
Wolfgang Guba, Agnes Meyder, Matthias Rarey, and Jérôme Hert
Journal of Chemical Information and Modeling 2016 56 (1), 1-5


interactive, desolvation-aware visual ΔG estimates

Task: Approximate and visualize affinities. The values are estimated based on a difference calculation between the bound and unbound state, based on an atomic logP-based mathematical kernel. The system has NOT been trained to specific targets, instead H-bond contribution and dehydration ("desolvation") are intrinsically balanced without weighting parameters as seen in all force fields. By design, HYDE allows the visualization of ΔG on atoms so that the user instantly gets a feedback on the computational details behind for lead optimization (L.O.) and other design tasks.
HYDE is constantly improved and originated from a collaboration with BAYER, Hamburg University, and BioSolveIT.

Towards an Integrated Description of Hydrogen Bonding and Dehydration: Decreasing False Positives in Virtual Screening with the HYDE Scoring Function
Ingo Reulecke, Gudrun Lange, Jürgen Albrecht, Robert Klein, and Matthias Rarey
Chem. Med. Chem. 2008 3 (6), 885-897
A consistent description of HYdrogen bond and DEhydration energies in protein–ligand complexes: methods behind the HYDE scoring function
Nadine Schneider, Gurdun Lange, Sally Hindle, Robert Klein, and Matthias Rarey
Journal of Computer-Aided Molecular Design 2013 27 (1), 15-29

pocket detection

druggable binding sites from 3D structures

Task: Compute proposals for "empty pockets"/binding sites, and visualize the results in 3D for further selection. The "Find empty sites" functionality is based on a heuristic model that employs Gaussian differences on a 3D grid to assess the likelihood of dealing with a pocket shape or not; besides this, global hydrogen bond functionalities and the lipophilic character plus the solvent accessible surface (SAS) of the putative pocket are taken into account. The computation is further enriched with local measures such as distances between pairs of functional group atoms.
The pocket detection algorithms are part of the DoGSiteScorer that emerged from a collaboration with Merck and Hamburg University.

DoGSiteScorer: a web server for automatic binding site prediction, analysis and druggability assessment
Andrea Volkamer, Daniel Kuhn, Friedrich Rippmann, and Matthias Rarey
Bioinformatics 2012 28 (15), 2074-2075
Analyzing the Topology of Active Sites: On the Prediction of Pockets and Subpockets
Andrea Volkamer, Axel Griewel, Thomas Grombacher, and Matthias Rarey
Journal of Chemical Information and Modeling 2010 50 (11), 2041-2052

FlexX docking

fast, flexible placement of ligands into cavities

Task: Place a ligand into a receptor cavity. The docking algorithm is a state-of-the-art fragment docker: Ligands are split into so-called fragments, and an initial fragment (or combinations thereof) are placed into multiple places in the pocket – and scored using a simple yet very fast pre-scoring scheme. From the n solutions placed, the ligand is further built up, fragment by fragment, and the interim solutions are scored against each other. The best scored survive the process, and those are delivered to the user. The initial idea relates back to the FlexX algorithm, however, many improvements have been made over the years, such as the "Single Interaction Scan" (SIS) placement that also finds solutions when there are only very few polar groups in a compound. The SIS algorithm uses virtual lines between protein and ligand interaction spots, and is trimmed for a very speedy rotation around these lines, thus generating solutions quickly for ligands of more hydrophobic character.

Placement of medium-sized molecular fragments into active sites of proteins
Matthias Rarey, Stephan Wefing, and Thomas Lengauer
Journal of Computer-Aided Molecular Design 1996 10 (1), 41-54
Substantial improvements in large-scale redocking and screening using the novel HYDE scoring function
Nadine Schneider, Sally Hindle, Gudrun Lange, Robert Klein, Jürgen Albrecht, Hans Briem, Kristin Beyer, Holger Claußen, Marcus Gastreich, Christian Lemmen, and Matthias Rarey
Journal of Computer-Aided Molecular Design 2012 26 (6), 701-723


stay tuned with all the breaking news.


© 2017 BioSolveIT GmbH,   all rights reserved