Improving Molecular-Morphology Contrastive Pretraining using Deep-Learning-based Morphology Profiles
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.
Related
- AET5: A transcriptome-guided molecular generation framework with contrastive self-supervised learning
- Katalyst: knowledge-guided semantic alignment via contrastive learning for enzyme turnover prediction
- EvoGraph: synergizing evolutionary semantics and tail-aware graph contrastive learning for protein-protein interaction prediction.
- Morphology, molecular phylogeny and azaspiracid profile of Azadinium poporum (Dinophyceae) from the Gulf of Mexico
- meth-SemiCancer2: a cancer subtyping framework leveraging domain adaptation with contrastive learning for batch effect correction in DNA methylation profiles
BibTeX
@misc{li2026improving,
title = {{Improving Molecular-Morphology Contrastive Pretraining using Deep-Learning-based Morphology Profiles}},
author = {Jie Li and Kathryn E. Kirchoff and Dante A. Pertusi and Zhizhuo Zhang},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.30433},
eprint = {2609.30433},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.30433},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Improving Molecular-Morphology Contrastive Pretraining using Deep-Learning-based Morphology Profiles
AU - Jie Li
AU - Kathryn E. Kirchoff
AU - Dante A. Pertusi
AU - Zhizhuo Zhang
PY - 2026
DA - 2026/09/24
PB - arXiv
DO - 10.48550/arxiv.2609.30433
UR - https://arxiv.org/abs/2609.30433
AB - 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.
ER -