Learning Cell-Aware Hierarchical Multi-Modal Representations for Robust Molecular Modeling
This work addresses the need for reliable molecular representations in biomedical modeling, offering a generalizable framework, though it appears incremental by building on existing cell-aware methods.
The paper tackled the problem of robust molecular property prediction by addressing limitations in current cell-aware approaches, such as modality incompleteness and insufficient hierarchical modeling, and proposed CHMR, a framework that improved classification by 3.6% and regression by 17.2% on average across nine benchmarks.
Understanding how chemical perturbations propagate through biological systems is essential for robust molecular property prediction. While most existing methods focus on chemical structures alone, recent advances highlight the crucial role of cellular responses such as morphology and gene expression in shaping drug effects. However, current cell-aware approaches face two key limitations: (1) modality incompleteness in external biological data, and (2) insufficient modeling of hierarchical dependencies across molecular, cellular, and genomic levels. We propose CHMR (Cell-aware Hierarchical Multi-modal Representations), a robust framework that jointly models local-global dependencies between molecules and cellular responses and captures latent biological hierarchies via a novel tree-structured vector quantization module. Evaluated on nine public benchmarks spanning 728 tasks, CHMR outperforms state-of-the-art baselines, yielding average improvements of 3.6% on classification and 17.2% on regression tasks. These results demonstrate the advantage of hierarchy-aware, multimodal learning for reliable and biologically grounded molecular representations, offering a generalizable framework for integrative biomedical modeling. The code is in https://github.com/limengran98/CHMR.