Abstract
One of the challenges facing efforts to generate molecular biomarkers for toxins is distinguishing between markers that are indicative of exposure and markers that provide evidence of the effects of toxicity. Phenotypic anchoring provides an approach to help segregate markers into these categories based on some phenotypic index of toxicity. Here we leveraged the mussel embryo-larval toxicity assay in which toxicity is estimated by the fraction of larvae that exhibit an abnormal morphology, to isolate subsets of larvae that were abnormal and thus showed evidence of copper-toxicity, versus others that while exposed to copper exhibited normal morphology. Mussel larvae reared under control conditions or in the presence of increasing levels of copper (3-15 μg/L Cu2+) were physically sorted according to whether their morphology was normal or abnormal, and then profiled using RNAseq. Supervised differential expression analysis identified sets of genes whose differential expression was specific to the pools of abnormal larvae versus normal larvae, providing putative markers of copper toxicity versus exposure. Markers of copper exposure and copper-induced abnormality were involved in many of the same pathways, including development, shell formation, cell adhesion, and oxidative stress, yet unique markers were detected in each gene set. Markers of effect appeared to be more resolving between phenotypes at the lower copper concentration, while markers of exposure were informative at both copper concentrations.