Abstract: AbstractIn this work we present an integrated system partly based on the commercially available software TOPKAT, which predicts chronic toxicity through provision of a computational estimation of Lowest Observed Adverse Effect Level (LOAEL) values. We found evidence that the LOAEL correlated with a specific class of molecular descriptors, known as 2D autocorrelation descriptors. The system developed is found to be helpful in supporting – with reasonable confidence – the prioritisation of issues in chemical food research, by establishing lev...
(read more)
Topics: 
Data mining
Machine learning
Statistics