7:30 Dinner (venue to be confirmed)
|9:45||Welcome||Rob Scoffin, Cresset|
|Protein-Ligand Interactions (Chair: Rob Scoffin, Cresset)|
|10:00||Polyphony: Superposition Independent Methods for Ensemble-Based Drug Discovery||Will Pitt, UCB|
|10:30||Under the Hood: Dealing with Protein Structures for Molecular Design||Matthias Rarey, University of Hamburg|
|11:15||Break and posters|
|11:45||Summarizing Activity and Selectivity in One Picture||Mark Mackey, Cresset|
|12:15||Generation of Toxicity Alerts and Solutions from Public Data Using Matched Molecular Pair Analysis (MMPA)||Lauren Reid, MedChemica|
|12:45||History of Imperial War Museum Duxford||David Lee, IWM Duxford|
|1:00||Lunch and posters|
|2:00||Toward an Understanding of the GPCR-ligand Interactions||Alex Heifetz, Evotec|
|2:30||Big is Not Best – Maximum Common Substructure-based Similarity Searching||Ed Duesbury, University of Sheffield|
|3:00||Break and posters|
|3:30||MOARF, an Integrated Workflow for Multiobjective Optimization||Nick Firth, Institute of Cancer Research|
|4:00||A Novel Docking Methodology for Membrane Proteins||Kiyoyuki Omoto, Pfizer Neusentis|
|4:30||Closing remarks and poster prize|
|5:00||Close | Delegates are free to view the museum exhibits until 6:00|
Will Pitt, UCB
UCB is targeting allosteric binding sites. As well as optimising ligand binding affinity, we would like to be able to increase efficacy by design. This is a difficult challenge for CADD scientists and one that requires that the dynamics of the protein target are taken into account. In this talk a methodology and open source software will be presented which is designed to analyse multiple protein-ligand crystal structures simultaneously. The aim is what we call ensemble-based drug discovery (EBDD). Published results and some information on how this approach is being applied to a current project will be described.
Matthias Rarey, University of Hamburg
Protein structures, mostly from X-ray crystallography, are the starting point for every structure-based molecular design endeavor. While highly precise structural information is necessary for modeling protein-ligand interactions, the experimental data regularly lacks details mostly resulting from missing hydrogen information. Furthermore, the structural model contained in a PDB file is an interpretation of electron density giving atomic detail a limited and varying reliability. Eventually, the question arises which structure should be taken and how the information from multiple structures can be ideally combined to consider important aspects of protein flexibility. In the talk, new algorithmic approaches for these problems will be surveyed. At work, they allow a more reliable and convenient work with protein structures for medicinal chemists – once the hood is closed.
Mark Mackey, Cresset
We have recently presented a method of summarizing the information obtained from 3D activity cliff analysis: examination of all pairs of molecules can distinguish between apparent cliffs that are outliers, or due to measurement error, and those which consistently point to particular electrostatic and steric features having a large impact on activity. To do this it has proved essential to allow for alignment noise: no 3D alignment technique is perfect, so we apply a Bayesian analysis to correct for potential misalignments and for the case where a molecule is aligned correctly except for a flexible substituent whose conformation is under-constrained. We use the recent AZ/CCDC alignment validation data set to determine valid estimates for the Bayesian priors.
As an extension of this technique, it is possible to mine the data for a simple picture of explored pharmacophoric space, corrected for the conformational and alignment flexibility of each molecule. This provides an invaluable picture to the chemist of which parts of property space around a molecule have been adequately explored. When considering a new molecule for synthesis, it is possible to compute the amount that this would increase the explored pharmacophoric space and hence present an ‘information content’ score for the new molecule: if we made and tested this new molecule, how much would it actually increase the SAR information content of the data set?
The combination of this, with the activity cliff summary data, allows a simple qualitative evaluation of the SAR of a data set in 3D, alongside guidance on which parts of pharmacophoric space have been mined out and which remain underexplored. We present the application of these techniques to several literature data sets.
Lauren Reid, MedChemica
The aim of this project was to extract safety knowledge, in the form of transformation rules and structural alerts, from public binding data to a selection of recommended safety assays. Tools developed by MedChemica were used to perform matched molecular pair analysis on curated ChEMBL18 binding data. As a result, over 12,000 transformation rules from 13 safety assays were generated, along with a list of related structural alerts. The methods used and issues addressed will be discussed.
1) Bowes, J. et al. Reducing safety-related drug attrition: the use of in vitro pharmacological profiling. Nature Reviews: Drug Discovery. 2012, 11(12), pp.909-922.
Drs. Alexander Heifetz1*, Ewa Chudyk1, Laura Gleave1, Mike Bodkin1, Dmitri G. Fedorov2, Philip C. Biggin3, Barrie Martin4 and Thorsten Nowak4
1Evotec (UK) Ltd., 114 Milton Park, Abingdon, Oxfordshire OX14 4SA, UK *firstname.lastname@example.org
2National Institute of Advanced Industrial Science and Technology (AIST), Central 2 Umezoto 1-1-1, Tsukuba, 305-8568 Japan
3Department of Biochemistry, University of Oxford, South Parks Road, Oxford OX1 3QU, UK
4C4X Discovery, Unit 310 Ducie House, Ducie Street, Manchester, M1 2JW, UK
GPCRs have enormous physiological importance, being the primary targets of a large number of modern drugs. The understanding of the binding interactions between the GPCR and the small molecule plays a key role in the rationalization of receptor-ligand affinity and are essential for the drug design process. We present here two conceptually pioneering approaches: 1) Fragment molecular orbitals (FMO) – quantum mechanical method and 2) NMR-based GPCR modelling protocol named HGMP-C4XD. These approaches allow exploration of GPCR ligand interactions and create a cost-efficient new avenue for structure-based drug discovery (SBDD) against GPCR targets.
Recent breakthrough in GPCRs crystallography resulted in the solving of >100 GPCR-ligand complexes. The ‘visual inspection’ and molecular mechanics (MM) calculations that are traditionally used for the rationalization of the receptor-ligand affinity cannot fully explain complex molecular interactions. The FMO method was adapted for accurate analysis of GPCR-ligand interactions and provides valuable insights into their chemical nature.
The hierarchical GPCR modeling protocol (HGMP) has been developed by Evotec in conjunction with the Oxford University to support GPCR SBDD programs. NMR-based C4XD technology was developed by C4X Discovery Ltd to explore how molecules behave in physiological-relevant solution. The combination of HGMP and the C4XD approaches allows the identification of the ligand’s bioactive conformation and for the accurate modeling of GPCR-ligand complex.
We are grateful to the Royal Society for an Industry Award granted to A.H (IF100104).
E. Duesbury, J. Holliday, P. Willett, University of Sheffield, Sheffield, UK
Finding the Maximum Common Substructure (MCS) of two molecules is an NP-Complete problem but the extent of the computation depends on how one defines the MCS. The connected MCS (cMCS), the largest fragment in common between two molecules, is far easier to calculate than the disconnected MCS (dMCS), i.e., the set of fragments yielding the largest overlap. The calculation of the dMCS can often be simplified by means of heuristics that simplify the search tree whilst minimising impact on the size of the MCS.1 In addition, one may impose topological distance constraints (tdMCS) to further simplify the search tree.2
This talk reviews the use of several different MCS algorithms when calculating the cMCS, dMCS and tdMCS for use in graph-based similarity searching. Our experiments used an open source virtual screening dataset3 , and have compared searching using the MCS with group fusion of extended connectivity fingerprints. The algorithms used fit into several categories, including clique detection algorithms, subgraph enumeration algorithms, spectral algorithms and similarity-based build-up algorithms. Some of these were exact methods, whereas others calculate an approximate MCS.
The conclusions from this investigation are as follows:
(1) Raymond, J. W.; Gardiner, E. J.; Willett, P. Heuristics for Similarity Searching of Chemical Graphs Using a Maximum Common Edge Subgraph Algorithm. J. Chem. Inf. Comput. Sci. 2002, 42, 305–316.
(2) Kawabata, T. Build-Up Algorithm for Atomic Correspondence between Chemical Structures. J. Chem. Inf. Model. 2011, 51, 1775–1787.
(3) Riniker, S.; Landrum, G. A. Open-Source Platform to Benchmark Fingerprints for Ligand-Based Virtual Screening. J. Cheminformatics 2013, 5, 26.
Nick Firth, Institute of Cancer Research
The aim of this work is to balance the drawbacks of de novo design using fragment replacement approach. A fragmentation algorithm, Synthetic Disconnection Rules (SynDiR), has been developed to enrich the design of synthetically tractable solutions. To permit the greatest possible coverage of chemical space we have designed and implemented a fragment replacement protocol, including a novel pharmacophoric fingerprint for Rapid Alignment of Topological Scaffolds (RATS). SynDiR and RATS have been integrated into a de novo design workflow, Multi Objective Automated Replacement of Fragments (MOARF), to enable the design of synthetically tractable molecules for multiobjective optimisation.
Kiyoyuki Omoto, Pfizer Neusentis
Recent progress in the X-ray crystallography has enabled the structure-based drug design (SBDD) across a variety of proteins. However, membrane proteins like ion channel are less liable to crystallization and homology modelling is necessary to enable the SBDD. A critical drawback of the homology modelling is low confidence in prediction of residue orientations in a binding site. Thus, a simple docking against a homology model is less likely successful. To overcome this gap, we have created a method ‘ensemble docking’ in which >1000 homology models are created and docking calculations are carried out for all the homology models. The huge number of docking poses provided are subjected to analyses such as clustering and interaction energy calculation to rank them with regard to likeliness to be a real binding mode. Of note is that this approach accounts for binding site flexibility as residue orientations are diverse across the homology models. In the presentation we will show some successful applications.
Permission has been given for the following presentations to be published.
|Summarizing Activity and Selectivity in One Picture||Mark Mackey, Cresset|
|Generation of Toxicity Alerts and Solutions from Public Data Using Matched Molecular Pair Analysis (MMPA)||Lauren Reid, MedChemica|
|Toward an Understanding of the GPCR-ligand Interactions||Alex Heifetz, Evotec|
|Big is Not Best – Maximum Common Substructure-based Similarity Searching||Ed Duesbury, University of Sheffield|
|MOARF, an Integrated Workflow for Multiobjective Optimization||Nick Firth, Institute of Cancer Research|
UK-QSAR autumn 2015 delegates
IWM Duxford is just south of Cambridge at Junction 10 of the M11 motorway, 30 miles from Junction 27 of the M25. It is also easily accessible from the A1, A14, M1 and the North, via the M11.
Follow signs to the AirSpace Conference centre. There is ample free parking and the museum car park can be used if required.
The closest stations are Whittlesford Parkway, Royston and Cambridge. There are taxi ranks at both Royston and Cambridge stations, however, Whittlesford Parkway does not have a taxi rank. Pre-booked taxis can be arranged – please indicate this when you register.
Cambridge to Whittlesford Parkway
8:48 arriving 8:57
London Liverpool Street to Whittlesford Parkway
7:28 arriving 8:43
7:58 arriving 9:13
London Kings Cross to Royston
7:52 arriving 8:41
8:05 arriving 9:07
8:10 arriving 9:07
London Kings Cross to Cambridge
07:44 arriving 08:33
7:52 arriving 9:04
08:14 arriving 09:04