ATTNSOM: Learning Cross-Isoform Attention for Cytochrome P450 Site-of-Metabolism

TL;DR

ATTNSOM is a novel deep learning framework that leverages cross-isoform attention mechanisms to improve the accuracy of predicting cytochrome P450 metabolic sites at the atom level, addressing limitations of previous models.

Contribution

The paper introduces ATTNSOM, a cross-isoform attention-based model that explicitly captures relationships across cytochrome P450 isoforms for enhanced site-of-metabolism prediction.

Findings

Achieves strong top-k performance across multiple isoforms

Yields higher Matthews correlation coefficient than ablated variants

Demonstrates the importance of modeling cross-isoform relationships

Abstract

Identifying metabolic sites where cytochrome P450 enzymes metabolize small-molecule drugs is essential for drug discovery. Although existing computational approaches have been proposed for site-of-metabolism prediction, they typically ignore cytochrome P450 isoform identity or model isoforms independently, thereby failing to fully capture inherent cross-isoform metabolic patterns. In addition, prior evaluations often rely on top-k metrics, where false positive atoms may be included among the top predictions, underscoring the need for complementary metrics that more directly assess binary atom-level discrimination under severe class imbalance. We propose ATTNSOM, an atom-level site-of-metabolism prediction framework that integrates intrinsic molecular reactivity with cross-isoform relationships. The model combines a shared graph encoder, molecule-conditioned atom representations, and a cross-attention mechanism to capture correlated metabolic patterns across cytochrome P450 isoforms. The model is evaluated on two benchmark datasets annotated with site-of-metabolism labels at atom resolution. Across these benchmarks, the model achieves consistently strong top-k performance across multiple cytochrome P450 isoforms. Relative to ablated variants, the model yields higher Matthews correlation coefficient, indicating improved discrimination of true metabolic sites. These results support the importance of explicitly modeling cross-isoform relationships for site-of-metabolism prediction. The code and datasets are available at https://github.com/dmis-lab/ATTNSOM.