A systematic investigation of molecular encoding methods for drug property predictions across neural network and Transformer encoder-based model
2026-06-08 • Machine Learning
Machine Learning
AI summaryⓘ
The authors studied how different ways of turning molecules into computer-readable formats affect how well models can predict molecular properties like toxicity and side effects. They tested two kinds of neural network models using several molecular fingerprint types on seven datasets and found high accuracy. Instead of using outside tools to explain their models, they looked inside the model's attention mechanisms to see which molecular features mattered. Their approach helped identify chemical groups important for properties like blood-brain barrier permeability and mutagenicity. This work offers advice on picking molecular encodings and builds toward more understandable drug discovery tools.
Molecular encodingNeural networksFingerprintsTransformer encoderBlood-brain barrier permeabilityMutagenicityAttention mechanismAUCInterpretabilityDrug discovery
Authors
Sheng-Ya Chen, Shan-Ju Yeh
Abstract
Fundamental investigations into how different molecular encoding methods affect molecular property prediction remain relatively limited. In this study, we extensively examined the optimal molecular encoding methods for molecular properties prediction using two prevalent structure designs: a classical neural network model (MLP) and a Transformer encoder-based model (MLP+TL). For molecular encoding methods, we investigated several types of fingerprints, including traditional topological fingerprints, substructure-based fingerprints, and string-based representations. These two models were trained on seven well-known molecular datasets to evaluate different input molecular encoding methods based on evaluation metrics. On several biologically relevant classification tasks, including toxicity, mutagenicity, and side-effect prediction, our models consistently achieved average AUC values above 0.9. Rather than relying on external post-hoc explanation methods such as the local interpretable model-agnostic explanation (LIME) or the Deep SHapley Additive exPlanations (SHAP), we leveraged the model's intrinsic attention weights as an internal interpretability signal for identifying potentially important feature. The MLP+TL model using MACCS and PubChem as input can capture chemically interpretable groups that determined the major blood-brain barrier (BBB) permeability and mutagenicity in Salmonella typhimurium. In particular, a comparison between Morphine and Heroin highlighted the role of hydroxyl-related substructures in BBB permeability prediction, which was consistently reflected in the attention weights. Overall, our findings provide practical guidance for selecting effective molecular encoding methods and contribute to the development of interpretable molecular informatics approaches for drug discovery.