ChemCLIP: Bridging Organic and Inorganic Anticancer Compounds Through Contrastive Learning

TL;DR

ChemCLIP introduces a contrastive learning framework that unifies organic and inorganic anticancer compounds into a shared representation space, enhancing cross-domain knowledge transfer and aiding drug discovery.

Contribution

It is the first to apply contrastive learning to bridge organic and inorganic anticancer compounds, creating a shared embedding space based on biological activity.

Findings

Morgan fingerprints achieved an average alignment ratio of 0.899.

Downstream classification AUCs were 0.859 for inorganic and 0.817 for organic compounds.

Contrastive learning effectively unifies chemically distinct compounds based on activity.

Abstract

The discovery of anticancer therapeutics has traditionally treated organic small molecules and metal-based coordination complexes as separate chemical domains, limiting knowledge transfer despite their shared biological objectives. This disparity is particularly pronounced in available data, with extensive screening databases for organic compounds compared to only a few thousand characterized metal complexes. Here, we introduce ChemCLIP, a dual-encoder contrastive learning framework that bridges this organic-inorganic divide by learning unified representations based on shared anticancer activities rather than structural similarity. We compiled complementary datasets comprising 44,854 unique organic compounds and 5,164 unique metal complexes, standardized across 60 cancer cell lines. By training parallel encoders with activity-aware hard negative mining, we mapped structurally distinct compounds into a shared 256-dimensional embedding space where biologically similar compounds cluster together regardless of chemical class. We systematically evaluated four molecular encoding strategies: Morgan fingerprints, ChemBERTa, MolFormer, and Chemprop, through quantitative alignment metrics, embedding visualizations, and downstream classification tasks. Morgan fingerprints achieved superior performance with an average alignment ratio of 0.899 and downstream classification AUCs of 0.859 (inorganic) and 0.817 (organic). This work establishes contrastive learning as an effective strategy for unifying disparate chemical domains and provides empirical guidance for encoder selection in multi-modal chemistry applications, with implications extending beyond anticancer drug discovery to any scenario requiring cross-domain chemical knowledge transfer.