Screening Level Estimation of Chemical Mixtures Toxicity Using In Silico Models: A JP-5 Case Study for Environmental Incidents

Environmental incidents can release multiple chemicals simultaneously, as seen in the recent JP-5 leak in Hawaii, raising concerns about human exposure through contaminated drinking water. Assessing the toxicity of such mixtures is essential, yet experimental data are often unavailable. This study evaluates the feasibility of applying publicly available in silico tools, including TTC, QSAR models, and read-across, to estimate oral points of departure (PODs) for acute, subchronic, and chronic exposures to JP-5. Applicability domain assessments indicated broad QSAR model coverage, with OPERA and ToxTree fully inclusive and VEGA models showing mostly moderate-to-high confidence, aside from a few additives. Dosimetric adjustments and uncertainty factors were applied to extrapolate rat PODs to in silico-derived health guidance values (HGV). Mixture toxicity was estimated by using the concentration addition model and compared with existing HGV-based values to derive screening levels. In silico-derived HGVs generally aligned with available guidance values, and the TTC estimates were consistently conservative. Potential health effects were also curated based on available animal and mechanistic studies. Overall, this case study demonstrates that in silico models can help fill toxicological data gaps and provide screening-level insights into evaluating chemical mixture exposure when empirical data are limited.