Modelling of Indicator Escherichia coli Contamination in Sentinel Oysters and Estuarine Water.

This study was performed to improve the ability to predict the concentrations of Escherichia coli in oyster meat and estuarine waters by using environmental parameters, and microbiological and heavy metal contamination from shellfish growing area in southern Thailand. Oyster meat (n = 144) and estuarine waters (n = 96) were tested for microbiological and heavy metal contamination from March 2016 to February 2017. Prevalence and mean concentrations of E. coli were 93.1% and 4.6 × 103 most probable number (MPN)/g in oyster meat, and 78.1% and 2.2 × 102 MPN/100 mL in estuarine water. Average 7-day precipitation, ambient air temperature, and the presence of Salmonella were associated with the concentrations of E. coli in oyster meat (p < 0.05). Raw data (MPN/g of oyster meat and MPN/100 mL of estuarine water) and log-transformed data (logMPN/g of oyster meat and logMPN/100 mL of estuarine water) of E. coli concentrations were examined within two contrasting regression models. However, the more valid predictions were conducted using non-log transformed values. These findings indicate that non-log transformed data can be used for building more accurate statistical models in microbiological food safety, and that significant environmental parameters can be used as a part of a rapid warning system to predict levels of E. coli before harvesting oysters.