Integrating chemical structures as treatments improves representations of microscopy images for morphological profiling
PLOS Computational Biology · 29.07.2026 · 22 (7) · e1014483
DOI: 10.1371/journal.pcbi.1014483 ↗Kurz gefasst
Fachartikel in der Zeitschrift PLOS Computational Biology, online erschienen am 29.07.2026. Verfasst von Yemin Yu und 7 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science, City University of Hong Kong, Kowloon Tong, Hong Kong. Hauptthemen laut MEDLINE: Microscopy, Image Processing, Computer-Assisted. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 42 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Recent advances in self-supervised deep learning have improved our ability to quantify cellular morphological changes in high-throughput microscopy screens, a process known as morphological profiling. However, most current methods only learn from images, despite many screens being inherently multimodal, as they involve both a chemical or genetic perturbation as well as an image-based readout. We hypothesized that incorporating chemical compound structures during self-supervised pre-training could improve learned representations of images from high-throughput microscopy screens. We introduce a representation learning framework, MICON (Molecular-Image Contrastive Learning), that models chemical compounds as treatments that induce transformations of cell phenotypes. MICON significantly outperforms classical hand-crafted features such as CellProfiler and existing deep-learning-based representation learning methods in challenging evaluation settings where models must identify reproducible effects of drugs across independent replicates and data-generating centers. We demonstrate that incorporating chemical compound information into the learning process provides small, but consistent improvements in performance and that modeling compounds specifically as treatments outperforms approaches that directly align images and compounds in a single representation space. Our findings point to a new direction for representation learning in morphological profiling, suggesting that methods should explicitly account for the multimodal nature of microscopy screening data.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- Department of Computer Science, City University of Hong Kong, Kowloon Tong, Hong Kong.
- Microsoft Research, Cambridge, Boston, Massachusetts, United States of America.
- Department of Computer Science, Zhejiang University, Hangzhou, China.
ORCID-Profile: Lester Mackey, Alex X Lu
Zitieren
APA
Yu, Y., Hayir, E., Tenenholtz, N., Mackey, L., Wei, Y., Alvarez-Melis, D., Amini, A. P., & Lu, A. X. (2026). Integrating chemical structures as treatments improves representations of microscopy images for morphological profiling. PLOS Computational Biology, 22(7), e1014483. https://doi.org/10.1371/journal.pcbi.1014483
Vancouver
Yu Y, Hayir E, Tenenholtz N, Mackey L, Wei Y, Alvarez-Melis D, et al. Integrating chemical structures as treatments improves representations of microscopy images for morphological profiling. PLOS Computational Biology. 2026;22(7):e1014483. doi:10.1371/journal.pcbi.1014483
BibTeX
@article{Yu2026,
author = {Yu, Yemin and Hayir, Emre and Tenenholtz, Neil and Mackey, Lester and Wei, Ying and Alvarez-Melis, David and Amini, Ava P and Lu, Alex X},
title = {Integrating chemical structures as treatments improves representations of microscopy images for morphological profiling},
journal = {PLOS Computational Biology},
year = {2026},
volume = {22},
number = {7},
pages = {e1014483},
doi = {10.1371/journal.pcbi.1014483}
}
EUVIMED