Improved molecular prediction using deep learning on cell images
Improving Molecular-Morphology Contrastive Pretraining using Deep-Learning-based Morphology Profiles
Machine Learning
Summary
Scientists developed a better way to predict how molecules affect cells by teaching computers to understand detailed cell images. They trained a method to match the behavior of molecules with changes in cell shapes using advanced AI techniques. This new method produces more informative data, helping predict things like drug toxicity more accurately. The method gets better with more training data and competes well with other models that use different biological information.
What this means in practice
- •For pharmaceutical data scientists: Generate molecular features based on cell image data to improve drug toxicity and activity prediction models.
- •For drug discovery chemoinformatics teams: Enhance QSAR prediction pipelines by integrating deep-learning-based morphological profiles derived from cell images.
Authors
Jie Li, Kathryn E. Kirchoff, Dante A. Pertusi, Zhizhuo Zhang
Abstract
Recent advancements in image-based profiling techniques have enabled the collection of high-volume cell morphology data, allowing new molecular embedding models to learn from the experimental phenotypic perturbations of a molecule in a cell. Previously, we developed Molecule-Morphology Contrastive Pretraining (MoCoP), a strategy for aligning small molecule embeddings to morphology fingerprints extracted through CellProfiler. The resulting molecular representation showed transferable performance for quantitative structure--activity relationship (QSAR) prediction tasks. Here, we extend the method by using a deep-learning-based cell image encoding pipeline to extract more feature-rich morphology profiles and align them to the molecular embeddings through contrastive learning. The new embeddings encode more accurate information on how molecules perturb cell morphology and enable improvements for QSAR predictions through either fixed-embedding linear probes or fully flexible fine-tuning. Morphology retrieval performance scales log-linearly with training data size, suggesting continued improvements as larger datasets become available. The improved MoCoP v2 also achieves superior performance on toxicity prediction and competitive results on ADME and activity benchmarks, when compared with existing molecular embedding models that use both cell morphology and transcriptomic data during training.