Recent technological advancements have led to the development of new high-throughput profiling methods, such as transcriptomics, that can be used to rapidly screen chemicals for potential hazards. Decreasing costs have made it feasible to profile all protein-coding genes across thousands of samples, allowing for broad evaluation of many target pathways and modes of action in a single screening assay. Similarly, it is now possible to apply high-content imaging across many different chemical exposures to capture a variety of changes in cell morphology. Such methods have been applied to in vitro chemical screening studies, including screening studies at EPA that were recently released on the CompTox Chemicals Dashboard. This type of data can be used for both hazard prediction and potency estimation, thereby informing risk assessments and prioritizing chemicals for further testing.
Assessing the reliability and reproducibility of these screening platforms is critical to their utility in regulatory applications. While these platforms often have lower signal-to-noise compared to individual targeted assays, the resulting data is also high-dimensional, allowing for the analysis of consistent trends across many molecular endpoints. This talk will provide an overview of computational methods and best practices for reliable analysis of high-throughput profiling data in a variety of use cases, and highlight the recent data release on the CompTox Chemicals Dashboard.