Prediction of skin sensitization potency using machine learning approaches. Issue 7 (10th January 2017)

• Record Type: Journal Article
• Authors: Zang, Qingda; Paris, Michael; Lehmann, David M.; Bell, Shannon; Kleinstreuer, Nicole; Allen, David; Matheson, Joanna; Jacobs, Abigail; Casey, Warren; Strickland, Judy
• Is Part Of: Journal of applied toxicology. Volume 37:Issue 7(2017)
• Page Start: 792
• Page End: 805
• Publication Date: 2017-01-10
• Subjects: Skin sensitization potency -- allergic contact dermatitis (ACD) -- integrated decision strategy (IDS) -- machine learning -- murine local lymph node assay (LLNA) -- direct peptide reactivity assay (DPRA) -- KeratinoSens -- h‐CLAT (human cell line activation test); Toxicology -- Periodicals; Industrial toxicology -- Periodicals; Environmentally induced diseases -- Periodicals; Toxicology -- Periodicals; 615.9005
• DOI: 10.1002/jat.3424
• Languages: English
• ISSNs: 0260-437X

Abstract

Abstract: The replacement of animal use in testing for regulatory classification of skin sensitizers is a priority for US federal agencies that use data from such testing. Machine learning models that classify substances as sensitizers or non‐sensitizers without using animal data have been developed and evaluated. Because some regulatory agencies require that sensitizers be further classified into potency categories, we developed statistical models to predict skin sensitization potency for murine local lymph node assay (LLNA) and human outcomes. Input variables for our models included six physicochemical properties and data from three non‐animal test methods: direct peptide reactivity assay; human cell line activation test; and KeratinoSens™ assay. Models were built to predict three potency categories using four machine learning approaches and were validated using external test sets and leave‐one‐out cross‐validation. A one‐tiered strategy modeled all three categories of response together while a two‐tiered strategy modeled sensitizer/non‐sensitizer responses and then classified the sensitizers as strong or weak sensitizers. The two‐tiered model using the support vector machine with all assay and physicochemical data inputs provided the best performance, yielding accuracy of 88% for prediction of LLNA outcomes (120 substances) and 81% for prediction of human test outcomes (87 substances). The best one‐tiered model predicted LLNA outcomes with 78% accuracy and human outcomes with 75% accuracy. By comparison, the LLNA predicts human potency categories with 69% accuracy (60 of 87 substances correctly categorized). These results suggest that computational models using non‐animal methods may provide valuable information for assessing skin sensitization potency. Copyright © 2017 John Wiley & Sons, Ltd. Abstract : Replacing animal testing for regulatory classification of skin sensitizers is a high priority for US federal agencies. We developed machine learning approaches combining data from three non‐animal methods with physicochemical property data to predict three categories of skin sensitization. The top model surpassed the accuracy of a standard animal test for skin sensitization, suggesting that computational models using data from non‐animal methods can provide valuable information to assess skin sensitization potency. Supporting Information.