Abstract
Of 47 endocrine-disrupting chemicals (EDCs) collected from literature and related to breast cancer, not all were tested in a toxicity forecaster (ToxCast) program of the US-Environmental Protection Agency (EPA). Therefore, deep learning models based on the toxicity data in that database have been used to predict the molecular toxicity of the untested EDCs. Combined with the values of median lethal doses (LDs), six potential targets of EDCs related to breast cancer have been identified, viz. MYC proto-oncogene, urokinase plasminogen activator receptor (PLAUR), cytochrome P450 4 A 11, nuclear receptor 1 H 2 (NR1H2), peroxisome proliferator-activated receptor alpha (PPARA), and hypoxia-inducible factor 1 alpha (HIF1A).
Author contributions
The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Supporting information
Supporting Information file includes the description of deep learning models based on the ToxCast assays (Table S1), the activity of EDCs for each assays using ToxCast in vitro assays and deep learning models based on ToxCast data (Table S2), and the potential assays related to breast cancer risk analyzed by point-biserial correlation (Table S3).