Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
May 6, 2026Asian Journal of Research in Computer ScienceOpen Access

EGAT-Pool: A Hierarchical Graph Attention Network with Edge-aware SAG Pooling for Robust Environmental Toxicity Prediction

View Full Paper
Ask AI
Bookmark
Share

Authors

PMPeter MakieuEDEdison D. DartueKYKwaku Oppong Yeboah-Amankwah

Discussion

Loading...

Member takes

Implication

A new GNN model achieves a Matthews Correlation Coefficient of 0.4068, indicating its effectiveness for toxicity prediction in chemical compounds and environmental risk assessment.

Key Points

  • The research aims to enhance environmental toxicity prediction using a novel GNN architecture, EGAT-Pool.
  • Developed EGAT-Pool integrating GATv2 and edge-aware SAG Pooling for molecular representation.
  • Evaluated the model on the Tox21 benchmark with a Murcko scaffold split protocol.
  • Utilized spatial attention heatmaps for interpretability.
  • EGAT-Pool achieved a Matthews Correlation Coefficient of 0.4068 on the Tox21 benchmark.
  • Model demonstrated a ROC-AUC of 0.8153, outperforming established GNNs like GCN, GAT, and GIN.
  • Identified known toxicophores autonomously, providing mechanistic interpretability.

Cite This Study

Makieu et al. (2026) studied this question.

synapsesocial.com/papers/69fa8eca04f884e66b5313abhttps://doi.org/10.9734/ajrcos/2026/v19i5858
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Environmental Assessment of Chemicals: Artificial Intelligence for Predicting Persistence, Bioaccumulation, and Toxicity Properties2026
  2. 2Adaptive Self-Attention Graph Pooling for Drug–Target Affinity Prediction2026
  3. 3Predicting Toxicity toward Nitrifiers by Attention-Enhanced Graph Neural Networks and Transfer Learning from Baseline Toxicity2025
  4. 4ADFC ‐ ATP : Attention‐Guided Dual‐View Fusion and Contrastive Pretraining for Robust Aquatic Toxicity Prediction2026 · 1 citations
  5. 5GraphADT: empowering interpretable predictions of acute dermal toxicity with multi-view graph pooling and structure remapping2024 · 16 citations