Cytometry masked autoencoder: An accurate and interpretable automated immunophenotyper
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.
Authors
- Sang‐Hyuk Jung (ORCID: https://orcid.org/0000-0003-4116-3327)
- Ajinkya Pattekar (ORCID: https://orcid.org/0000-0002-1416-1230)
- Patryk Orzechowski (ORCID: https://orcid.org/0000-0003-3578-9809)
- Mark M. Painter (ORCID: https://orcid.org/0000-0002-0180-2748)
- Jakob Woerner
- C.A.G. Ittner
- Sheetal Gandotra (ORCID: https://orcid.org/0000-0002-5475-9704)
- Lamorna Brown Swigart (ORCID: https://orcid.org/0000-0003-2076-5177)
- Matei Ionita (ORCID: https://orcid.org/0000-0002-0514-3364)
- Jaesik Kim (ORCID: https://orcid.org/0000-0003-1203-9650)
- Jeffrey McKeehan
- Divij Mathew (ORCID: https://orcid.org/0000-0002-8323-7358)
- Kevin W. Gibbs (ORCID: https://orcid.org/0000-0002-5272-2289)
- Nuala J. Meyer (ORCID: https://orcid.org/0000-0003-4597-5584)
- Joost Wagenaar (ORCID: https://orcid.org/0000-0003-0837-7120)
- Yonghyun Nam (ORCID: https://orcid.org/0000-0003-2963-2151)
- Nilam S. Mangalmurti (ORCID: https://orcid.org/0000-0002-9146-1608)
- E. John Wherry (ORCID: https://orcid.org/0000-0003-0477-1956)
- Carolyn S. Calfee (ORCID: https://orcid.org/0000-0001-9208-6865)
- Ellen Lyda Burnham (ORCID: https://orcid.org/0000-0002-9945-0284)
- Harsh Barot
- Kathleen D. Liu (ORCID: https://orcid.org/0000-0003-4891-8649)
- Allison R. Greenplate (ORCID: https://orcid.org/0000-0002-2614-3072)
- Dokyoon Kim (ORCID: https://orcid.org/0000-0002-4592-9564)
- Cynthia Clendenin (ORCID: https://orcid.org/0000-0003-4535-2088)
- Sokratis A. Apostolidis (ORCID: https://orcid.org/0000-0002-0224-0097)
- Michael A. Matthy
- Matthew Lee (ORCID: https://orcid.org/0000-0002-2842-7191)
- John P. Reilly (ORCID: https://orcid.org/0000-0003-3937-5320)
- Alexandra P. Turner (ORCID: https://orcid.org/0000-0003-3577-8587)
- Derek W. Russel
- Karl W. Thomas
- Michelle L. McKeague
- MIKA ESPERANZA
- Van Truong (ORCID: https://orcid.org/0009-0007-3118-7946)
- Thomas Dunn
- Dylan T. Norton
Institutions
- University of California, San Francisco (US)
- University of Alabama at Birmingham (US)
- Parker Institute for Cancer Immunotherapy (US)
- Birmingham VA Medical Center (US)
- Translational Therapeutics (United States) (US)
- Wake Forest University (US)
- AGH University of Krakow (PL)
- University of San Francisco (US)
- University of Pennsylvania (US)
- University of Colorado Denver (US)
Publication Details
- Journal
- Cell Reports Medicine
- Published
- 2024-11-01
- DOI
- https://doi.org/10.1016/j.xcrm.2024.101808
- Citations
- 4
- Primary Topic
- Single-cell and spatial transcriptomics
- Type
- article
- Field-Weighted Citation Impact
- 0.94
Funders
- University of Pennsylvania
- National Institutes of Health
- Perelman School of Medicine, University of Pennsylvania