🤖 AI Summary
To address high model complexity, poor generalizability, and excessive computational cost in drug–drug interaction (DDI) prediction, this work proposes a lightweight neural network–based minimalist modeling paradigm. Methodologically, we systematically evaluate three molecular representations—Morgan fingerprints, GCN-based graph embeddings, and MoLFormer molecular transformer embeddings—and construct compact fully connected models under leak-proof data splitting. We further integrate Grad-CAM–style gradient analysis for interpretability validation. Key contributions include: (1) demonstrating that simple molecular representations achieve state-of-the-art performance (AUC > 0.92 on DrugBank/FDA datasets) even under stringent generalization settings; (2) the first identification—via interpretable analysis—of clinically relevant pharmacophores, including CYP inhibition and P-glycoprotein substrate motifs; and (3) challenging the “complexity bias” in DDI modeling by advocating a new paradigm prioritizing data quality and incremental model development over architectural sophistication.
📝 Abstract
Accurately predicting drug-drug interactions (DDIs) is crucial for pharmaceutical research and clinical safety. Recent deep learning models often suffer from high computational costs and limited generalization across datasets. In this study, we investigate a simpler yet effective approach using molecular representations such as Morgan fingerprints (MFPS), graph-based embeddings from graph convolutional networks (GCNs), and transformer-derived embeddings from MoLFormer integrated into a straightforward neural network. We benchmark our implementation on DrugBank DDI splits and a drug-drug affinity (DDA) dataset from the Food and Drug Administration. MFPS along with MoLFormer and GCN representations achieve competitive performance across tasks, even in the more challenging leak-proof split, highlighting the sufficiency of simple molecular representations. Moreover, we are able to identify key molecular motifs and structural patterns relevant to drug interactions via gradient-based analyses using the representations under study. Despite these results, dataset limitations such as insufficient chemical diversity, limited dataset size, and inconsistent labeling impact robust evaluation and challenge the need for more complex approaches. Our work provides a meaningful baseline and emphasizes the need for better dataset curation and progressive complexity scaling.