Background: Phenotypic high-throughput screens are often utilized in modern oncology drug discovery pipelines. Such screens are conducted with an array of molecules (panels of oncology drugs, drugs for other indications, and compounds with possibly unknown bioactivity), ultimately measuring a biological change within a disease model. For example, a screen might test drugs in NF1 wild-type and mutant cancer cell lines to find molecules that are selectively toxic to NF1 mutant cells. While these screens can yield valuable hits, they also present the challenge of identifying the target(s) that mediate the effect seen in the screen for several reasons. For example, many drugs have a polypharmacologic target profile making it hard to identify which targets are biologically relevant for a given drug. This can cause a single drug to be treated as a specific inhibitor of different targets by different researchers. Furthermore, some libraries contain molecules for which drug-target studies have never been conducted. To address these challenges, we developed an application that enables exploration of the chemical-biological interaction space. Methods: Data from ChEMBL, PubChem, DrugBank, and the Drug-Gene Interaction Database (DGIdb) for over 7000 small molecules were downloaded. Drug-target interaction data points as well as evidence to the contrary were quantified for each interaction. In addition, when quantitative data were obtained, median IC50s were calculated. Each molecule was annotated with a name and chemical structure, and every target was annotated with protein & gene identifiers. To enable exploration of this database, an interactive web interface was developed using the R Shiny platform and cheminformatics R packages (rcdk, fingerprint, webchem).Results: The app enables the end user with a specific query molecule to search a database of experimentally-derived drug-target interactions. The database can be queried using drug names or structures. A similarity parameter allows the user to expand their search to other structurally related molecules. The structure of the query molecule is compared to every database molecule, and structurally similar molecules and targets are presented in interactive tabular and network-based forms for in-depth exploration. The app also performs enrichment analysis on the target lists, and allows the user to evaluate structure-activity relationships for drug response data using publicly-available in vitro datasets. Finally, if a user has a target of interest, they can search for molecules that bind that target and explore the resulting data interactively. Conclusions: This app provides a multifunctional platform for exploring chemical space as it relates to biological targets, and may be useful at several steps along the oncology drug development pipeline including target discovery, structure-activity relationship, and lead identification studies.

Citation Format: Robert J. Allaway, Sara J. Gosline, Marco Nievo, Salvatore La Rosa, Annette Bakker, Justin Guinney. Drug-Target Explorer: An interactive tool for examining chemical-biological interactions [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 4643.