Cytometry masked autoencoder: An accurate and interpretable automated immunophenotyper

• Published in: Cell reports. Medicine Vol. 5; no. 11; p. 101808
• Main Authors: Kim, Jaesik, Ionita, Matei, Lee, Matthew, McKeague, Michelle L., Pattekar, Ajinkya, Painter, Mark M., Wagenaar, Joost, Truong, Van, Norton, Dylan T., Mathew, Divij, Nam, Yonghyun, Apostolidis, Sokratis A., Clendenin, Cynthia, Orzechowski, Patryk, Jung, Sang-Hyuk, Woerner, Jakob, Ittner, Caroline A.G., Turner, Alexandra P., Esperanza, Mika, Dunn, Thomas G., Mangalmurti, Nilam S., Reilly, John P., Meyer, Nuala J., Calfee, Carolyn S., Liu, Kathleen D., Matthy, Michael A., Swigart, Lamorna Brown, Burnham, Ellen L., McKeehan, Jeffrey, Gandotra, Sheetal, Russel, Derek W., Gibbs, Kevin W., Thomas, Karl W., Barot, Harsh, Greenplate, Allison R., Wherry, E. John, Kim, Dokyoon
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 19.11.2024; Elsevier
• Subjects: Advanced Basic Science; Algorithms; automated gating; deep learning; Flow Cytometry - methods; high-dimensional cytometry; Humans; immunophenotyping; Immunophenotyping - methods; machine learning; mass cytometry; representation learning; Single-Cell Analysis - methods; deep learning; automated gating; representation learning; immunophenotyping; high-dimensional cytometry; mass cytometry; machine learning
• ISSN: 2666-3791; 2666-3791
• DOI: 10.1016/j.xcrm.2024.101808

Single-cell cytometry data are crucial for understanding the role of the immune system in diseases and responses to treatment. However, traditional methods for annotating cytometry data face challenges in scalability, robustness, and accuracy. We propose a cytometry masked autoencoder (cyMAE), which automates immunophenotyping tasks including cell type annotation. The model upholds user-defined cell type definitions, facilitating interpretability and cross-study comparisons. The training of cyMAE has a self-supervised phase, which leverages large amounts of unlabeled data, followed by fine-tuning on specialized tasks using smaller amounts of annotated data. The cost of training a new model is amortized over repeated inferences on new datasets using the same panel. Through validation across multiple studies using the same panel, we demonstrate that cyMAE delivers accurate and interpretable cellular immunophenotyping and improves the prediction of subject-level metadata. This proof of concept marks a significant step forward for large-scale immunology studies. [Display omitted] •Masked cytometry modeling learns relationships among proteins without cell identity•We develop cytometry masked autoencoder (cyMAE) to automate immunophenotyping•cyMAE improves both cell-level and subject-level immune profiling Kim et al. introduces cyMAE, a cytometry masked autoencoder that automates immune cell profiling from single-cell cytometry data. By leveraging unlabeled data for pre-training and fine-tuning on specific tasks, the model improves immune profiling accuracy and enhances prediction of subject-level clinical data, advancing large-scale immune studies.