Poster 29: Dealing with Combinatorial Chemical Space: Towards a Universal FrameworkFlorian Lauck1, Matthias Rarey1
|1Center for Bioinformatics, University of Hamburg|
|The search for new drug molecules is challenging since chemical space is almost infinite. The collective strategy is to limit the search space so that it is easier to find only those molecules that have suitable physicochemical and topological properties. Therefore, one requires methods to efficiently model this chemical subspace avoiding its full enumeration. Chemical space can be modeled in a combinatorial fashion utilizing molecular fragments with respect to synthetic accessibility. Here, we present a universal framework for handling combinatorial chemical spaces, so called fragment spaces. |
A fragment space consists of molecular fragments and connection rules[1,2]. Each fragment has at least one reaction site, which corresponds to an open valence. These sites are modeled as dummy atoms called linkers. The connection rules determine compatibility of such linkers. When two fragments are connected, the dummy atoms are removed and a bond in accordance with the connection rule is introduced.
Connection rules are usually derived from known chemical reactions, but may be motivated differently. We differentiate between two strategies. In the knowledge-based approach small, experimentally accessible molecules are added as fragments while exchanging their functional groups with linkers. Synthetic reactions are used to derive connection rules. In the library-based approach drug-like molecules are cleaved into fragments[2,4]. Connection rules are derived from retrosynthetic analysis and subsequently used to identify bonds of a molecule suitable for cleaving. For each molecule these bonds are removed and linkers are added to the adjacent atoms. The resulting fragments are then added to the fragment space. While the latter approach can be automated to an extent, the former helps to design a synthetic protocol for a new molecule.
Fragment spaces are a rich source of new molecules since they can describe large - even infinite - chemical space in a very compact, space efficient format. However, enumerating all possible molecules in a brute-force manner most often results in combinatorial explosion. Therefore, retrieving molecules is non-trivial and requires sophisticated algorithms. No matter how a fragment space was compiled, the same tools can be applied for querying.
There exist several strategies for retrieving molecules from a fragment space. Although, a full enumeration is not feasible, enumeration with stringent physicochemical constraints is possible. It is also possible to first generate a focused fragment space describing a certain class of molecules, in other words, a smaller chemical space that may be fully enumerated. Furthermore, a fragment space can be searched based on a query molecule, the major problem being that new molecules may span over several fragments. Therefore, different molecular descriptors were adapted to the fragment space context. Another type of search retrieves molecules with a common substructure as defined by a SMARTS pattern[7,8]. Other applications of fragment spaces include structure based molecular design and scaffold replacement[10,11].
Our framework provides access to the aforementioned functionality. It allows for straightforward creation and manipulation of fragment spaces, as well as querying them. It was designed to efficiently deal with combinatorial chemical space and retrieve new molecules and sets thereof quickly with different search strategies. For the first time, we also present a graphical user interface for inspecting and manipulating fragment spaces interactively.
1. Rarey M, Stahl M (2001) Similarity searching in large combinatorial chemistry spaces. J Comput Aided Mol Des 15: 497520.
2. Lewell XQ, Judd DB, Watson SP, Hann MM (1998) RECAP-Retrosynthetic Combinatorial Analysis Procedure:  A Powerful New Technique for Identifying Privileged Molecular Fragments with Useful Applications in Combinatorial Chemistry. J Chem Inf Comp Sci 38: 511522. Available: http://pubs.acs.org/cgi-bin/doilookup/?10.1021/ci970429i.
3. Detering C, Claussen H, Gastreich M, Lemmen C (2010) KnowledgeSpace - a publicly available virtual chemistry space Vol. 2. p. O9. doi:10.1186/1758-2946-2-S1-O9.
4. Degen J, Wegscheid-Gerlach C, Zaliani A, Rarey M (2008) On the art of compiling and using drug-like chemical fragment spaces. ChemMedChem 3: 15031507. doi:10.1002/cmdc.200800178.
5. Pärn J, Degen J, Rarey M (2007) Exploring fragment spaces under multiple physicochemical constraints. J Comput Aided Mol Des 21: 327340. doi:10.1007/s10822-007-9121-3.
6. Fischer JR, Lessel U, Rarey M (2010) LoFT: Similarity-Driven Multiobjective Focused Library Design. J Chem Inf Model 50: 121. doi:10.1021/ci900287p.
7. Ehrlich H-C, Volkamer A, Rarey M (2012) Searching for Substructures in Fragment Spaces. J Chem Inf Model 0: null. doi:doi: 10.1021/ci300283a.
8. Ehrlich H-C, Henzler A, Rarey M (n.d.) Note for the editor: The paper will be submitted in January, the reference can be updated with the proofs of this book chapter. J Chem Inf Model.
9. Degen J, Rarey M (2006) FlexNovo: structure-based searching in large fragment spaces. ChemMedChem 1: 854868. doi:10.1002/cmdc.200500102.
10. Maaß P, Schulz-Gasch T, Stahl M, Rarey M (2007) Recore:  A Fast and Versatile Method for Scaffold Hopping Based on Small Molecule Crystal Structure Conformations. J Chem Inf Model 47: 390399. doi:10.1021/ci060094h.
11. Boehm M, Wu T-Y, Claussen H, Lemmen C (2008) Similarity Searching and Scaffold Hopping in Synthetically Accessible Combinatorial Chemistry Spaces. J Med Chem 51: 24682480. doi:10.1021/jm0707727.