ChiENN: Embracing Molecular Chirality with Graph Neural Networks

被引:1
作者
Gainski, Piotr [1 ]
Koziarski, Michal [2 ,3 ]
Tabor, Jacek [1 ]
Smieja, Marek [1 ]
机构
[1] Jagiellonian Univ, Fac Math & Comp Sci, Krakow, Poland
[2] Mila Quebec Inst, Montreal, PQ, Canada
[3] Univ Montreal, Montreal, PQ, Canada
来源
MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES: RESEARCH TRACK, ECML PKDD 2023, PT III | 2023年 / 14171卷
关键词
Graph Neural Networks; GNN; Message-passing; Chirality; Molecular Property Prediction;
D O I
10.1007/978-3-031-43418-1_3
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Graph Neural Networks (GNNs) play a fundamental role in many deep learning problems, in particular in cheminformatics. However, typical GNNs cannot capture the concept of chirality, which means they do not distinguish between the 3D graph of a chemical compound and its mirror image (enantiomer). The ability to distinguish between enantiomers is important especially in drug discovery because enantiomers can have very distinct biochemical properties. In this paper, we propose a theoretically justified message-passing scheme, which makes GNNs sensitive to the order of node neighbors. We apply that general concept in the context of molecular chirality to construct Chiral Edge Neural Network (ChiENN) layer which can be appended to any GNN model to enable chirality-awareness. Our experiments show that adding ChiENN layers to a GNN outperforms current state-of-the-art methods in chiral-sensitive molecular property prediction tasks.
引用
收藏
页码:36 / 52
页数:17
相关论文
共 50 条
  • [41] Graph Unrolling Networks: Interpretable Neural Networks for Graph Signal Denoising
    Chen, Siheng
    Eldar, Yonina C.
    Zhao, Lingxiao
    IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2021, 69 : 3699 - 3713
  • [42] Graph Neural Networks Pretraining Through Inherent Supervision for Molecular Property Prediction
    Benjamin, Roy
    Singer, Uriel
    Radinsky, Kira
    PROCEEDINGS OF THE 31ST ACM INTERNATIONAL CONFERENCE ON INFORMATION AND KNOWLEDGE MANAGEMENT, CIKM 2022, 2022, : 2903 - 2912
  • [43] Node-aware convolution in Graph Neural Networks for Predicting molecular properties
    Linh Le Pham Van
    Quang Bach Tran
    Tien Lam Pham
    Quoc Long Tran
    2020 12TH INTERNATIONAL CONFERENCE ON KNOWLEDGE AND SYSTEMS ENGINEERING (IEEE KSE 2020), 2020, : 120 - 125
  • [44] Heterogeneous graph neural networks with denoising for graph embeddings
    Dong, Xinrui
    Zhang, Yijia
    Pang, Kuo
    Chen, Fei
    Lu, Mingyu
    KNOWLEDGE-BASED SYSTEMS, 2022, 238
  • [45] GDLL: A Scalable and Share Nothing Architecture Based Distributed Graph Neural Networks Framework
    Duong Thi Thu Van
    Khan, Muhammad Numan
    Afridi, Tariq Habib
    Ullah, Irfan
    Alam, Aftab
    Lee, Young-Koo
    IEEE ACCESS, 2022, 10 : 21684 - 21700
  • [46] Scalable Graph Neural Networks with Deep Graph Library
    Zheng, Da
    Wang, Minjie
    Gan, Quan
    Song, Xiang
    Zhang, Zheng
    Karypis, Geroge
    WSDM '21: PROCEEDINGS OF THE 14TH ACM INTERNATIONAL CONFERENCE ON WEB SEARCH AND DATA MINING, 2021, : 1141 - 1142
  • [47] Scalable Graph Neural Networks with Deep Graph Library
    Zheng, Da
    Wang, Minjie
    Gan, Quan
    Zhang, Zheng
    Karypis, Geroge
    KDD '20: PROCEEDINGS OF THE 26TH ACM SIGKDD INTERNATIONAL CONFERENCE ON KNOWLEDGE DISCOVERY & DATA MINING, 2020, : 3521 - 3522
  • [48] Learning Graph Neural Networks with Deep Graph Library
    Zheng, Da
    Wang, Minjie
    Gan, Quan
    Zhang, Zheng
    Karypis, George
    WWW'20: COMPANION PROCEEDINGS OF THE WEB CONFERENCE 2020, 2020, : 305 - 306
  • [49] Imperceptible graph injection attack on graph neural networks
    Yang Chen
    Zhonglin Ye
    Zhaoyang Wang
    Haixing Zhao
    Complex & Intelligent Systems, 2024, 10 : 869 - 883
  • [50] Imperceptible graph injection attack on graph neural networks
    Chen, Yang
    Ye, Zhonglin
    Wang, Zhaoyang
    Zhao, Haixing
    COMPLEX & INTELLIGENT SYSTEMS, 2024, 10 (01) : 869 - 883