You are here:
Relationships between aquatic toxicity, chemical hydrophobicity and mode of action: log kow QSARs revisited
Citation:
Barron, M., D. Vivian, AND C. Stevens. Relationships between aquatic toxicity, chemical hydrophobicity and mode of action: log kow QSARs revisited. SETAC Europe, Rome, N/A, ITALY, May 13 - 17, 2018.
Impact/Purpose:
This abstract will be a presentation at the SETAC Europe science conference. The impact and value to EPA include: 1) the work represents a significant advancement in the analysis of aquatic toxicity-chemical hydrophobicity relationships, which have been the basis for most aquatic toxicity quantitative structure activity relationships (QSARs); 2) EPA relies on these QSARs as a fundamental step in the TSCA hazard assessment, and are the basis of EPA's ECOSAR program; and 3) by developing and assessing the value of developing mode of action (MOA)-specific models, the research provides important information for the Agency to determine the next generation of QSARs.
Description:
Relationships between chemical hydrophobicity and toxicity have been shown for nearly 100 years in both mammals and fish, typically using the log of the octanol:water partition coefficient (kow). The current study reassessed the influence of mode of action (MOA) on aquatic toxicity-log kow relationships using a comprehensive database of curated and standardized acute toxicity and consensus log Kow values, and weight of evidence MOA classifications. Log Kow models were developed as linear regressions of log acute toxicity and log kow for 50 different combinations of taxa (e.g., fish, invertebrates, species-specific) and MOA (6 broad; 3 specific narcosis subtypes). MOA categories included narcosis (non-polar, polar, ester), acetylcholinesterase inhibition, neurotoxicity, electron transport inhibition, iono/osmoregulatory/circulatory impairment, and reactivity. Forty-eight of the 50 MOA-based models were statistically significant (p<0.05; most p<0.001), but r2 values were generally less than 0.5, particularly for non-narcosis MOAs. The results showed that MOA-based models can improve the accuracy of aquatic toxicity predictions for a range of taxa, and that incorrect classification of a specific acting chemical can result in toxicity prediction errors greater than 1000 fold.