You are here:
Multivariate Models for Prediction of Human Skin Sensitization Hazard
Citation:
Strickland, J., Q. Zang, M. Paris, N. Kleinstreuer, D. Lehmann, D. Allen, N. Choksi, J. Matheson, A. Jacobs, A. Lowit, AND W. Casey. Multivariate Models for Prediction of Human Skin Sensitization Hazard. Future Tox III, Arlington, VA, November 19 - 20, 2015.
Impact/Purpose:
One of ICCVAM’s top priorities is the development and evaluation of non-animal approaches to identify potential skin sensitizers. The complexity of biological events necessary for a substance to elicit a skin sensitization reaction suggests that no single alternative method will replace animal tests. The performance of the integrated approach presented herein was better than the local lymph node assay or any of the in chemico, in vitro, or in silico methods alone. These data suggest that computational methods are promising tools to effectively identify potential skin sensitizers without testing animals.
Description:
One of ICCVAM’s top priorities is the development and evaluation of non-animal approaches to identify potential skin sensitizers. The complexity of biological events necessary for a substance to elicit a skin sensitization reaction suggests that no single alternative method will replace animal tests. Data from the direct peptide reactivity assay (DPRA), human cell line activation test (h-CLAT), KeratinoSens assay, six physicochemical properties, and an in silico read-across prediction of skin sensitization hazard served as inputs to two machine learning approaches. Support vector machine and logistic regression were used to integrate the data, which were applied in 12 feature set combinations, to predict human skin sensitization hazard. Models were trained on a set of 72 substances and tested on an external set of 24 substances. The feature set containing DPRA, h-CLAT, KeratinoSens, read-across, and log P performed the best for both approaches: accuracy = 99% (71/72), sensitivity = 98% (50/51), and specificity = 100% (21/21) for the training set; and accuracy = 96% (23/24), sensitivity = 93% (14/15), and specificity = 100% (9/9) for the test set. The performance of this integrated approach was better than the local lymph node assay or any of the in chemico, in vitro, or in silico methods alone. These data suggest that computational methods are promising tools to effectively identify potential skin sensitizers without testing animals. This project was funded in whole or in part with Federal funds from the NIEHS, NIH under Contract No. HHSN273201500010C.