Update on the ToxCast Pipeline and Invitrodb
The US Environmental Protection Agency (EPA) ToxCast data pipeline (tcpl) has been applied to hundreds of assay endpoints covering diverse biological endpoints, including approximately 1800 genes and bioactivities including mitochondrial toxicity, nuclear receptor activation, immune response, and developmental toxicity, among others. As open source R package software, tcpl enables first tier data processing of heterogeneous bioactivity data from high-throughput screening (HTS), including curve-fitting and estimation of potency values. The primary objective of this presentation is to describe major updates to tcpl and the database, invitrodb, to increase assay annotation, update functions for cytotoxicity threshold estimation, expand available curve-fitting models, and enhance plotting capabilities. The number of publicly available assay endpoints in the upcoming version 4 release has increased to nearly 600 assays, 1300 components, and 1900 endpoints, with updates in assay annotations to enable better aggregation. The lower bound estimate of the cytotoxicity threshold has been updated, requiring a 5% positive rate in curated cytotoxicity assays (now 88). Tcpl curve-fitting models (constant, Hill, and gain-loss models) have been expanded to include Polynomial 1 (Linear), Polynomial 2 (Quadratic), Power, Exponential 2, Exponential 3, Exponential 4, and Exponential 5 based on BMDExpress and encoded by R package dependency tcplFit2. Inclusion of these models impacted invitrodb in two primary ways: (1) the need for long-format storage of generic modeling parameters; and, (2) the need for updated logic on selection of winning model. Continuous hitcall probability in tcplfit2 is calculated as the product of 3 different probabilities: median response and top of model exceed the cutoff, and the AIC is less than the constant model. Plotting in tcpl has been expanded with a new utility called tcplPlot. TcplPlot provides interactive display of concentration response curves through plotly integration and has built in REST api functionality. Integrating tcplPlot with tcpl will allow for a consistent visualization of curves in reports and web applications. Tcpl and invitrodb provide a standard for consistent and reproducible curve-fitting and data management for diverse in vitro assay data, with easily available documentation, thus enabling the sharing and use of these data in myriad toxicology applications. This abstract does not necessarily reflect U.S. EPA policy.
Impact/Purpose
The primary objective of this presentation is to describe major updates to tcpl and the database, invitrodb, to increase assay annotation, update functions for cytotoxicity threshold estimation, expand available curve-fitting models, and enhance plotting capabilities.Citation
Brown, J., M. Feshuk, K. Houck, R. Judson, AND K. Paul-Friedman. Update on the ToxCast Pipeline and Invitrodb. Society of Toxicology Annual Meeting (Virtual), N/A, N/A, March 12 - 26, 2021. [DOI: 10.23645/epacomptox.14681241]Download(s)
- BROWN_SOT_2021_FINAL.PDF (PDF) (NA pp, 538.5 KB, about PDF)