PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
Synapse
⌘+K
Synapse
May 7, 2026Journal of CheminformaticsOpen Access

Framework for evaluating explainable AI in antimicrobial drug discovery

View Full Paper
Ask AI
Bookmark
Share

Authors

AOAbdulmujeeb T. OnawoleMBMark A. T. BlaskovichJZJohannes Zuegg

Discussion

Loading...

Member takes

Overview

Framework assesses explainable AI's role in drug development, highlighting structure-activity relationships and model robustness.

Key Points

  • The aim is to develop an evaluation framework for explainable AI in antimicrobial drug discovery.
  • Developed evaluation framework using fragment-based explainability tests
  • Employed Random Forest, CNN, and RGCN models for molecular representation
  • Compared models on scaffold recognition and robustness
  • Analyzed explainability behaviour regarding activity cliffs
  • All XAI methods showed good predictive capabilities
  • Random Forest and molecular graphs performed well in scaffold recognition
  • Highlighting different explainability behaviours for activity cliffs among the XAI approaches

Cite This Study

Onawole et al. (2026) studied this question.

synapsesocial.com/papers/69fbf004164b5133a91a443ahttps://doi.org/10.1186/s13321-026-01200-x
View Full Paper
Ask AI
Bookmark
Share