The Lilly Reachable Chemical Space System: Bridging Chemical Synthesis Potential with Discovery ChemistryChristos A Nicolaou1, Ian Watson1, Jibo Wang1
|1Lilly Research Laboratories, Eli Lilly and Company|
|Efforts for mapping chemical space and reaching out to less explored, but potentially promising regions for pharmaceutical development have been hampered by the sheer number of theoretically feasible compounds and the practical concern on the synthesizability of the chemical structures proposed. In a typical setting, such virtual compounds are conceived through the enumeration of the products of chemical reactions when supplied reagent sets appropriate to the specific reaction. Alternatively, virtual compounds may be proposed simply through the permutation of a set of atoms abiding to some rudimentary chemical structure rules. In either case the result is a large virtual collection of chemical structures of questionable synthesizability and, therefore, practical use. |
This presentation provides a description of LiRCS, the Lilly Reachable Chemicals Space system, designed to bridge the chemical synthesis knowhow and potential at Eli Lilly with the needs of ongoing discovery chemistry projects. LiRCS can be thought of as the computational counterpart to our Automated Synthesis Lab (ASL) system which served as the main motivation for this work. In its current incarnation, the system focuses almost exclusively on reactions validated on the ASL.
LiRCS consists of the Lilly Annotated Reaction Repository (LLARR), a flexible virtual synthesis engine (VSE) supported by a high-performance computing system, and, a collection of search and retrieve utilities dedicated to a number of user-defined usage scenarios. In order to ensure synthesizability of the proposed compounds, LLARR relies on the usage of an annotated reaction scheme which captures a wealth of information on a reaction, including the detailed profile of the reagents that may (or may not) be used. Reactions in LLARR can be used by the VSE to either enumerate the full matrix of virtual compounds possible with a set of reagents or to simply generate structures by combining specific reagents subject to user imposed restrictions on the final products, related to e.g. compound size or chemical structure. VSE is tightly coupled with the Lilly chemical sample management system to ensure that the reagent sets used in the process can readily be accessible for chemical synthesis.
The system is designed to support a number of usage scenarios by Lilly scientists. Among them are the generation of virtual screening sets based on user defined objectives; the retrieval of reachable compounds satisfying structural similarity restrictions to a query compound (LiRCS-Grow); the retrieval of the synthetic route of a specific query compound (LiRCS-Retrieve). The presentation will provide an account of the algorithmic development that led to the system stressing the techniques implemented to manage the size of the reachable chemical space. Specific examples showcasing results produced by the system and demonstrating its capabilities will be presented. A discussion on lessons learned, issues to be resolved, and future development directions will conclude the presentation.