Multivariate Analysis and Prediction of Dioxin-Furan Congener (TEQ) Toxicity in Freshwater Fish based on Fatty Acid Methyl Ester (FAME) Profiles
Citation:
Multivariate Analysis and Prediction of Dioxin-Furan Congener (TEQ) Toxicity in Freshwater Fish based on Fatty Acid Methyl Ester (FAME) Profiles .
Impact/Purpose:
Dioxins, which are bioaccumulative and environmentally persistent, pose an ongoing risk to human and ecosystem health. Fish constitute a significant source of dioxin exposure for humans and fish-eating wildlife. Current dioxin analytical methods are costly, time-consuming, and produce hazardous by-products. A Danish team developed a novel, multivariate statistical methodology based on the covariance of dioxin-furan congener Toxic Equivalences (TEQs) and fatty acid methyl esters (FAMEs) and applied it to North Atlantic Ocean fishmeal samples. The goal of the current study was to attempt to extend this Danish methodology to 77 whole and composite fish samples from three trophic groups: predator (whole largemouth bass), benthic (whole flathead and channel catfish) and forage fish (composite bluegill, pumpkinseed and green sunfish) from two dioxin contaminated rivers (Pocatalico R. and Kanawha R.) in West Virginia, USA. Multivariate statistical analyses, including, Principal Components Analysis (PCA), Hierarchical Clustering, and Partial Least Squares Regression (PLS), were used to assess the relationship between the FAMEs and TEQs in these dioxin contaminated freshwater fish from the Kanawha and Pocatalico Rivers. These three multivariate statistical methods all confirm that the pattern of Fatty Acid Methyl Esters (FAMEs) in these freshwater fish covaries with and is predictive of the WHO TEQ 2005 toxicity, accomplishing the study goal. Fish Relative Weight (Wr), a measure of condition or well-being, was also compared with dioxin toxicity and it appears to decline with dioxin exposure. Multivariate statistical analytical techniques are powerful tools to reveal non-self-evident patterns and relationships in data. Given the various, high costs of acquiring and validating environmental data it is prudent to use the appropriate multivariate techniques to extract the maximal amount of information allowing us “to move rapidly from data to information to knowledge” (Smilde, Bro and Geladi 2004). (Note: draft final report will not be revised and completed; initial premise was found to be flawed. )
Description:
Peer Review Draft of Regional Methods Initiative Final Report