You are here:
Predictive Models of target organ and Systemic toxicities (BOSC)
Citation:
Shah, I., M. Martin, Chris Grulke, L. Pham, J. Liu, A. Williams, G. Patlewicz, AND R. Thomas. Predictive Models of target organ and Systemic toxicities (BOSC). Presented at CSS BOSC Meeting, RTP, NC, November 16 - 18, 2016. https://doi.org/10.23645/epacomptox.5180278
Impact/Purpose:
Poster presented at the BOSC CSS meeting in RTP, NC
Description:
The objective of this work is to predict the hazard classification and point of departure (PoD) of untested chemicals in repeat-dose animal testing studies. We used supervised machine learning to objectively evaluate the predictive accuracy of different classification and regression algorithms using chemical structure information, physico-chemical properties, and in vitro bioactivity data. The mean F1 score for predicting 20 target-organ hazard classes across three guideline study types was 0.69, and the R2 for predicting the PoD for systemic toxicity was 0.38. These models can be used to efficiently prioritize tens of thousands of environmental chemicals by hazard and by PoD.
URLs/Downloads:
DOI: Predictive Models of target organ and Systemic toxicities (BOSC)
CSS BOSC_PREDICTIVE_MODELS_POSTER_V2_JC.PDF (PDF, NA pp, 630.453 KB, about PDF)