From Nuclear Receptors to GPCRs: a Deep Transfer Learning Approach for Enhanced Environmental Estrogen Recognition.

in Environmental science & technology by Tingji Yao, Jiaqi Luo, Xiaoxiao Han, Hang Yi, Huazhou Zhang, Wenxiao Pan, Qiao Xue, Xian Liu, Jianjie Fu, Aiqian Zhang

TLDR

  • This study develops a deep transfer learning model, GPNET, to predict GPER-binding ligands using 3D molecular surface electrostatic potential point clouds as input.

Abstract

Environmental estrogens (EEs), as typical endocrine-disrupting chemicals (EDCs), can bind to classic estrogen receptors (ERs) to induce genomic effects, as well as to G protein-coupled estrogen receptor (GPER) located on the membrane, thereby inducing downstream nongenomic effects rapidly. However, due to the relatively scarce ligand data, receptor-based or ligand-based screening models for GPER are challenging. Inspired by functional similarity between GPER and ER, this study constructs a deep transfer learning model named GPNET to predict potential GPER-binding ligands by using three-dimensional (3D) molecular surface electrostatic potential point clouds (SepPC) as input. The model retains a part of molecular structural knowledge learned from the ER ligands and then trains the remaining parameters of the model using the GPER ligands, ultimately obtaining the GPNET model, which effectively predicts the binding activity of compounds with GPER. GPNET outperforms From Scratch (nontransfer) model on the small data set, achieving the area under the receiver operating characteristic (ROC) curve (AUC) of 0.898 on the validation set and 0.863 on the test set, respectively. Furthermore, by visualizing the critical points and extracting the features from activation points of active ligands, the study provides a more in-depth interpretation of the molecular mechanism of two bisphenol A (BPA) alternatives binding to GPER.

Overview

  • The study aims to predict potential GPER-binding ligands using a deep transfer learning model named GPNET.
  • The model uses three-dimensional molecular surface electrostatic potential point clouds (SepPC) as input and leverages the functional similarity between GPER and ER.
  • The primary objective is to develop a robust model for predicting GPER-binding ligands and to provide insights into the molecular mechanism of ligand-binding.

Comparative Analysis & Findings

  • The GPNET model outperforms the From Scratch (non-transfer) model on the small dataset, achieving an AUC of 0.898 on the validation set and 0.863 on the test set.
  • The study finds that GPNET is effective in predicting the binding activity of compounds with GPER.
  • The visualization and feature extraction of activation points of active ligands provide insights into the molecular mechanism of two bisphenol A (BPA) alternatives binding to GPER.

Implications and Future Directions

  • The study's findings demonstrate the potential of GPNET for predicting GPER-binding ligands, which can aid in the development of safer alternatives to BPA.
  • Future studies can focus on expanding the dataset and incorporating more molecular features to improve the model's accuracy.
  • The GPNET model can be applied to other GPER-mediated biological processes, advancing our understanding of the molecular mechanisms involved.
Read Full Article