From Cellular Responses to Pharmacological Domains: Multimodal Zero-Shot Drug Representation Learning

2026-07-28

AI Summary

Existing methods in multimodal drug representation learning struggle to disentangle mechanism-relevant signals from modality-specific noise, limiting their ability to predict pharmacological properties of unseen compounds. This work proposes the PMRD framework, which constructs a consensus pharmacological response space across chemical structures, gene expression profiles, and cellular morphologies to decouple mechanism-consistent factors from modality-specific information. PMRD incorporates a dynamic reweighting and reliability-aware multi-view retrieval mechanism, guided by pharmacological responses, enhanced with mechanistic candidate augmentation, and optimized via geometric attribution objectives. Evaluated on public benchmarks, the method substantially improves zero-shot prediction performance, generates drug neighborhoods with greater biological coherence, and effectively mitigates representational conflicts between structurally dissimilar yet pharmacologically similar compounds.

Abstract

Multimodal drug discovery enables drug representation learning beyond chemical structure by incorporating cellular responses such as gene expression and cell morphology. However, direct fusion and instance-level contrastive alignment may mix mechanism-related signals with modality-specific noise and incorrectly separate structurally dissimilar but biologically related compounds. This limitation can obscure transferable mechanism patterns required for predicting the properties of unseen compounds. We introduce PMRD, a pharmacological response domain-guided framework for multimodal zero-shot drug property prediction. PMRD separates mechanism-consistent factors from modality-specific information and constructs a consensus response domain across three modalities. Mechanism candidate augmentation identifies locally stable factors, while retrieval-geometry attribution dynamically reweights the alignment and augmentation objectives according to whether their updates preserve inter-drug discriminability.This feedback suppresses training signals that conflict with mechanism-discriminative retrieval. PMRD further combines complementary representations through reliability-aware multiview retrieval. Experiments on public datasets show improved zero-shot property prediction and more biologically coherent drug neighborhoods. Hard-negative analysis further indicates fewer conflicts between structurally dissimilar but response-related compounds. These results support PMRD as an effective framework for mechanism-aware multimodal drug representation learning.\footnote{The code will be released upon publication.}

Problem

• multimodal drug representation
• zero-shot drug property prediction
• mechanism-aware learning
• cellular response
• modality-specific noise

Innovation

• multimodal drug representation
• zero-shot learning
• pharmacological response domain
• mechanism-aware learning
• reliability-aware retrieval