Learning Molecular Representation in a Cell

2024/06/17 by Gang Liu, Srijit Seal, Liu, Gang +11 · 3 citations
Computer Science · Biochemistry, Genetics and Molecular Biology · Materials Science · #Computational Drug Discovery Methods #Cell Image Analysis Techniques #Machine Learning in Materials Science

paper · doi:10.48550/arxiv.2406.12056

Abstract

ment (InfoAlign) approach to learn molecular representations through the information bottleneck method in cells. We integrate molecules and cellular response data as nodes into a context graph, connecting them with weighted edges based on chemical, biological, and computational criteria. For each molecule in a training batch, InfoAlign optimizes the encoder's latent representation with a minimality objective to discard redundant structural information. A sufficiency objective decodes the representation to align with different feature spaces from the molecule's neighborhood in the context graph. We demonstrate that the proposed sufficiency objective for alignment is tighter than existing encoder-based contrastive methods. Empirically, we validate representations from InfoAlign in two downstream applications: molecular property prediction against up to 27 baseline methods across four datasets, plus zero-shot molecule-morphology matching.

Cited by

• From Cellular Responses to Pharmacological Domains: Multimodal Zero-Shot Drug Representation Learning
• Learning Cell-Aware Hierarchical Multi-Modal Representations for Robust Molecular Modeling
• Efficient Cell Painting Image Representation Learning via Cross-Well Aligned Masked Siamese Network

Related

• How Molecules Impact Cells: Unlocking Contrastive PhenoMolecular Retrieval
• Cross-modal representation alignment of molecular structure and perturbation-induced transcriptional profiles
• Removing Biases from Molecular Representations via Information Maximization
• Molecular Graph Representation Learning Integrating Large Language Models with Domain-specific Small Models
• Learning Cell-Aware Hierarchical Multi-Modal Representations for Robust Molecular Modeling