Drug discovery and development has become a resource-intensive process which is characterized by high attrition rates of candidate molecules. A deep characterization of candidate molecules during pre-clinical development requires efficient and cost-effective methods and the development of data rich assays. High-content imaging can provide a cost-effective solution to capture a broad range of biological responses. However, analysis of image data requires extensive expertise and data processing pipelines. These limitations and requirements, together with the lack of reference databases for phenotypic profiles, remain a major hurdle for drug development in routinely employing high-content analysis in drug discovery. With the REMATCH project we aid drug discovery by providing scalable software solutions and a unique reference knowledge base of compound profiles for phenotypic drug discovery. To eliminate bottlenecks in the processing of high volume of image data, we created a scalable, cloud-based software stack. In proof-of-concept drug screens, we then acquired millions of images of drug treated cancer cell lines. Using our improved, cloud-based analysis platform we demonstrated and benchmarked the analysis of high volume image data sets. This created a unique knowledge base of compound profiles with wide applicability in drug discovery. This cloud-based software and reference data sets has raised first interests in our solutions by academic partners and pharmaceutical companies. In summary, we demonstrated proof-of-concept of the cloud-based REMATCH software and we analyzed the currently largest data set of its’ kind. Further intellectual property protection of our approach is currently being explored.
