Development of expert alerts for toxicological endpoints is a time-consuming process, especially if changes are subsequently requested to these alerts based on proprietary data which for legal reasons cannot be shared with the original implementer even under strict confidence. In this poster we present a method for the acceleration of this process, analogous with the student-teacher model used for statistical methods in similar situations. In order to demonstrate this method, the endpoint of phototoxicity was chosen.
Validating proposed improvements to phototoxicity expert alerts against proprietary data
Ponting DJ; Payne M; Gerets H; Harvey JS; Howe J; Parenty A; White A; Williams RV; Macmillan DS;