Skin Doctor: Machine Learning Models for Skin Sensitization Prediction that Provide Estimates and Indicators of Prediction Reliability.

Jochen Kühnl,Anke Wilm,Andreas Schepky,Johannes Kirchmair,Christoph Bauer,Conrad Stork

doi:10.3390/ijms20194833

Abstract

The ability to predict the skin sensitization potential of small organic molecules is of high importance to the development and safe application of cosmetics, drugs and pesticides. One of the most widely accepted methods for predicting this hazard is the local lymph node assay (LLNA). The goal of this work was to develop in silico models for the prediction of the skin sensitization potential of small molecules that go beyond the state of the art, with larger LLNA data sets and, most importantly, a robust and intuitive definition of the applicability domain, paired with additional indicators of the reliability of predictions. We explored a large variety of molecular descriptors and fingerprints in combination with random forest and support vector machine classifiers. The most suitable models were tested on holdout data, on which they yielded competitive performance (Matthews correlation coefficients up to 0.52; accuracies up to 0.76; areas under the receiver operating characteristic curves up to 0.83). The most favorable models are available via a public web service that, in addition to predictions, provides assessments of the applicability domain and indicators of the reliability of the individual predictions.

Highlights

Repeated exposure to reactive chemicals with skin-sensitizing properties can cause allergic contact dermatitis (ACD) [1], an adverse cutaneous condition with a prevalence of ~20% among the general population [2] and even higher prevalence among workers with chronic occupational exposure [3]
In order to develop a detailed understanding of the relevance of the available local lymph node assay (LLNA) data to modeling the skin sensitization potential of xenobiotics, we analyzed the composition and molecular
Sci. 2019, 20, 4833 diversity of the LLNA data sets of Alves et al and Di et al In addition, we assessed how well the individual LLNA data sets cover the chemical space of cosmetics, approved drugs and pesticides

Summary

Introduction

Repeated exposure to reactive chemicals with skin-sensitizing properties can cause allergic contact dermatitis (ACD) [1], an adverse cutaneous condition with a prevalence of ~20% among the general population [2] and even higher prevalence among workers with chronic occupational exposure [3]. Understanding the skin sensitization potential of small organic molecules is of essence to the development and safe application of chemicals, including cosmetics and drugs. Animal tests have effectively been the only method for determining the skin sensitization potential and potency of substances. New in vitro and in chemico methods have been developed and evaluated [6,7,8,9], and computational approaches are starting to be recognized as important alternatives to animal testing [8,9,10,11]. The non-redundant combinatorial use of said methods in defined approaches that assess several key events of the adverse outcome pathway (AOP) for skin sensitization shows promising predictive capacity [12] and is currently evaluated in risk assessment case studies

Objectives

Methods

Results

Conclusion