Abstract
Crude oil toxicity studies in marine organisms largely focus on whole-mixture effects and a limited set of polycyclic aromatic hydrocarbons (PAHs), leaving the toxicity of many other oil-derived constituents poorly characterized. Here, we aimed to prioritise single oil-derived chemicals with high propensity for receptor-mediated activity and evaluate whether in silico selection translates to ex vivo transcriptional responses in Atlantic cod (Gadus morhua) liver, a model species in oil-related toxicology studies. We established a scalable workflow, using open source software, combining computational screening and experimental validation using precision-cut liver slices. A total of 159 identified constituents in a resin-like fraction of a crude oil water accommodated fraction (WAF) were screened against the aryl hydrocarbon receptor (Ahr2), estrogen receptor alpha (Era) and peroxisome proliferator-activated receptor alpha (Ppara) using three docking models. Pose stability was evaluated by molecular dynamics simulations, yielding 10 priority compounds. Liver slices were exposed to the prioritised compounds, along with selected positive controls. Preliminary results indicate that 5 of the 6 tested compounds elicited statistically significant transcriptional changes (RT-qPCR) in at least one gene target, with the strongest responses observed within the 10-100 µM range. These findings support mechanistically informed prioritisation as an efficient strategy to reduce large, data-poor chemical lists from complex mixtures to shortlists for targeted toxicological follow-up beyond traditionally monitored PAHs.