A strong consensus has emerged over the past 10 years in the pharmaceutical industry that the drug-target- focused drug discovery approach, to which the industry has subscribed almost entirely since the 1980s, has been a failure. Accordingly, over the past few years, the majority of R&D in the drug discovery and development space has been re-allocated to support alternative approaches, notably phenotypic screening to discover new drugs. This shift represents a new and growing business opportunity for new, big data science and technology products, such as those developed by the applicant organization, GeneCentrix. In common phenotypic screens, a chemical library is first screened against cells (e.g. HeLa cells), and a search is then undertaken for the primary molecular target of the most active compound. This search, known as target deconvolution, is challenging and is a rate-limiting step in the success of the screen. Accordingly, new technologies have been developed for target deconvolution. Specifically, target-specific arrays of drug-like chemical probes are increasingly being used in phenotypic screens (e.g. Novartis MOA box, the NIHs MIPE platform, etc.). GeneCentrix is pioneering the development of an information product that can enhance phenotypic screens that use these target-specific arrays. Specifically, the off-targets of the target-specific compounds (polypharmacology) in the array are currently not taken into account in analyzing the results of the phenotypic screen. In addition, the expression level of drug targets in the cell line used in the phenotype screen is very likely a predictive variable in target deconvolution but is also not taken into account. Therefore, we propose two aims to adapt our technology for the purpose of re- ranking phenotype screen results: 1) generation of off-target annotation (polypharmacologic profiles) for all compounds in a standard target-specific array; 2) integration of the polypharmacologic profile with expression data of the targets, and re-ranking of targets by this combined score. We will then validate the performance of our technology versus standard analysis of screening results; for a test set of phenotypic screens, an improvement of 20% in sensitivity associated with an equal or greater accuracy, as measured by area under the receiver operating curve (AUC), will be considered sufficient validation to pursue technical and commercial feasibility of the product in a Phase II application.
Public Health Relevance Statement: Project Narrative Big pharma early drug discovery is moving towards use of phenotypic discovery methods (as opposed to target- specific methods), and towards use of target-specific compounds in phenotypic screening. Integration of polypharmacologic target affinity data and gene expression data for those targets into the output of phenotypic screens that use target-specific compound libraries for target deconvolution will enhance the value of these screens, allowing for more accurate deconvolution of the mechanism of action of chemical probes emerging from phenotypic screens. This project thus harnesses extensive and diverse big biomedical data and translates it into knowledge of sufficient integrity to be packaged as a viable commercial product.
Project Terms: Affinity; Algorithms; Analytical Chemistry; Area; arm; base; big biomedical data; Big Data; Big Data to Knowledge; Biological Markers; Businesses; CDK2 gene; Cell Line; Cells; Chemical Actions; Chemicals; Collection; Consensus; Custom; Data; Data Science; Databases; Development; Docking; drug development; drug discovery; Drug Industry; Drug Targeting; drug testing; Ecology; Economics; Engineering; Exhibits; Failure; Gene Expression; Generations; Gold; Hela Cells; Image; improved; Industry; Informatics; inhibitor/antagonist; Investments; Knowledge; Libraries; Ligands; Literature; Measures; Methods; Modeling; Molecular Computations; Molecular Mechanisms of Action; Molecular Target; Nature; new technology; novel therapeutics; Output; Participant; Pathway interactions; Performance; Pharmaceutical Preparations; Pharmacologic Substance; Phase; Phenotype; Probability; Race; Research; research and development; ROC Curve; Scanning; screening; Screening Result; small molecule; small molecule libraries; success; Technology; Testing; Tissues; Translating; United States National Institutes of Health; Validation
