A major component of the National Toxicology Program’s effort to evolve the practice of toxicology from an observational science to one that is more mechanistic involves using functional omic technologies, such as transcriptomics. NTP’s approach to genomic dose-response modeling focuses on identifying biological potency of the test article. The approach reflects recommendations of experts in toxicology and toxicogenomics. The steps in the process include 1) developing a study design that includes a broad range of doses that sufficiently describe the shape of the dose-response and maximizes the accuracy of benchmark dose (BMD) values derived for all responsive genes; 2) incorporating a signal detection filter to ensure the modeled data contain adequate signal to lend confidence that any effect is treatment related; 3) conducting effect size and trend tests to identify those genes exhibiting a biologically plausible and reproducible response to test article treatment; 4) fitting of parametric dose-response models derived from the EPA BMD software to identify a biological potency estimate (i.e., a BMD) for each gene exhibiting a dose-related response to treatment; 5) grouping of genes into predefined Gene Ontologies (GO), identifying the GO gene sets responding to treatment, and then determining the composite potency of the gene set (i.e., a gene-set BMD); and 6) providing a biologically and mechanistically focused interpretation. NTP has developed a software package, BMDExpress 2.0, which enables users to implement the approach and perform extensive data visualization. The approach to genomic dose-response modeling is anticipated to lead to greater consistency in reporting of genomic dose-response data and to facilitate the use of genomic dose-response data in risk assessment.
National Toxicology Program. 2018. NTP research report on National Toxicology Program approach to genomic dose-response modeling. NTP RR 5. Research Triangle Park, NC. National Toxicology Program (5) 1-44. https://doi.org/10.22427/NTP-RR-5