Feature Space Selection and Heterogeneous Effect Estimation for Blood-Brain Barrier Permeability: A Random Forest to the Generalized Random Forest Pipeline

2026-09-24

AI Summary

This study addresses the challenge of blood-brain barrier (BBB) permeability prediction in central nervous system drug discovery by investigating heterogeneous causal relationships between molecular structures and permeability. Using the MoleculeNet BBBP dataset, we systematically ablate feature spaces to decouple representation from architectural effects. Modeling employs dynamic and generalized random forests alongside double machine learning, incorporating Morgan fingerprints and RDKit descriptors. A dynamic random forest with combined features achieves an optimal AUC of 0.970. The core contribution lies in elucidating the coupling mechanisms between features and algorithms, and demonstrating that, following orthogonalization, the LogP-BBB association exhibits no significant systematic heterogeneity, thereby correcting the overestimation bias inherent in naive causal forests.

Abstract

Predicting blood-brain barrier (BBB) permeability is critical for central nervous system drug discovery. Using the MoleculeNet BBBP dataset (n = 2039), this study systematically ablates molecular feature spaces to isolate featurisation from model architecture. We evaluate three feature families (Morgan fingerprints, RDKit physicochemical descriptors, SMILES bigrams) across four learning algorithms. Results demonstrate that predictive performance depends jointly on feature representation and algorithm. Dynamic Random Forest using combined features achieved the highest mean AUC (0.970, 95% CI: 0.963-0.977). Second, this optimal representation enables exploratory estimation of heterogeneous associations between molecular structure and BBB permeability using Generalized Random Forests. Constructing a pseudo-treatment from a LogP median split, we applied double/debiased machine learning to account for confounding. Orthogonalization substantially attenuates the heterogeneity detected by naive causal forests; no conditional effects remained significant after false discovery rate correction (smallest adjusted p = 0.082). Furthermore, orthogonalized feature importance shifted toward residual structural information in SMILES bigrams. Ultimately, once observed confounding is properly accounted for, evidence that LogP-BBB associations vary systematically across chemical space is insufficient. This underscores that feature representation and model architecture are coupled design choices, and that unorthogonalized causal forests risk overstating genuine treatment effect heterogeneity.

Problem

• Blood-Brain Barrier Permeability
• Feature Space Selection
• Heterogeneous Effect Estimation
• Confounding
• Causal Forest

Innovation

• Feature Space Selection
• Generalized Random Forests
• Double/Debiased Machine Learning
• Heterogeneous Effect Estimation
• Blood-Brain Barrier Permeability