한빛사
>
한빛사논문
한빛사논문
- 조회130
Cell Rep. Med., Nov 04 2024, S2666-3791(24)00555-X | https://doi.org/10.1016/j.xcrm.2024.101808
Cytometry masked autoencoder: An accurate and interpretable automated immunophenotyper
Jaesik Kim 1,2,3,21, Matei Ionita 2,5,21, Matthew Lee 2,3,4, Michelle L. McKeague 2,5, Ajinkya Pattekar 2,5, Mark M. Painter 2,5, Joost Wagenaar 2,4, Van Truong 2,3,4, Dylan T. Norton 2,5, Divij Mathew 2,5, Yonghyun Nam 2,3,4, Sokratis A. Apostolidis 2,6, Cynthia Clendenin 2, Patryk Orzechowski 2,4,7, Sang-Hyuk Jung 2,3,4, Jakob Woerner 2,3,4, Caroline A.G. Ittner 8, Alexandra P. Turner 8, Mika Esperanza 8, Thomas G. Dunn 9, Nilam S. Mangalmurti 2,8, John P. Reilly 2,8, Nuala J. Meyer 8, Carolyn S. Calfee 10,11,12, Kathleen D. Liu 13, Michael A. Matthy 12, Lamorna Brown Swigart 14, Ellen L. Burnham 15, Jeffrey McKeehan 15, Sheetal Gandotra 16, Derek W. Russel 16,17, Kevin W. Gibbs 18, Karl W. Thomas 18, Harsh Barot 19, Allison R. Greenplate 2,5, E. John Wherry 2,5,20, Dokyoon Kim 1,2,3,4,22
1Department of Bioengineering, University of Pennsylvania, Philadelphia, PA, USA
2Institute for Immunology & Immune Health (I3H), Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
3Institute for Biomedical Informatics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
4Department of Biostatistics, Epidemiology and Informatics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
5Department of Systems Pharmacology and Translational Therapeutics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
6Division of Rheumatology, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
7Department of Automatics and Robotics, AGH University of Science and Technology, al. Mickiewicza 30, 30-059 Kraków, Poland
8Division of Pulmonary and Critical Care Medicine, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
9Division of Hematology/Oncology, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
10Department of Anesthesia and Perioperative Care, University of California, San Francisco, School of Medicine, San Francisco, CA 94143, USA
11Division of Pulmonary, Critical Care, Allergy, and Sleep Medicine, University of California, San Francisco, School of Medicine, San Francisco, CA 94143, USA
12Cardiovascular Research Institute, Department of Medicine, University of California, San Francisco, School of Medicine, San Francisco, CA 94158, USA
13Division of Nephrology and Critical Care Medicine, University of California, San Francisco, School of Medicine, San Francisco, CA 94143, USA
14Department of Laboratory Medicine, University of California, San Francisco, School of Medicine, San Francisco, CA 94143, USA
15Division of Pulmonary Sciences and Critical Care Medicine, Department of Medicine, University of Colorado School of Medicine, Aurora, CO 80045, USA
16Division of Pulmonary, Allergy and Critical Care Medicine, Department of Medicine, University of Alabama at Birmingham, Birmingham, AL 35294, USA
17Pulmonary Section, Birmingham Veteran’s Affairs Medical Center, Birmingham, AL 35233, USA
18Section on Pulmonary and Critical Care, Allergy, and Immunology, Department of Internal Medicine, Wake Forest School of Medicine, Winston-Salem, NC 27157, USA
19Section on Hospital Medicine, Department of Internal Medicine, Wake Forest School of Medicine, Winston-Salem, NC 27157, USA
20Parker Institute for Cancer Immunotherapy, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
21These authors contributed equally
22Lead contact
Corresponding authors: Allison R. Greenplate, E. John Wherry, Dokyoon Kim
1Department of Bioengineering, University of Pennsylvania, Philadelphia, PA, USA
2Institute for Immunology & Immune Health (I3H), Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
3Institute for Biomedical Informatics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
4Department of Biostatistics, Epidemiology and Informatics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
5Department of Systems Pharmacology and Translational Therapeutics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
6Division of Rheumatology, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
7Department of Automatics and Robotics, AGH University of Science and Technology, al. Mickiewicza 30, 30-059 Kraków, Poland
8Division of Pulmonary and Critical Care Medicine, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
9Division of Hematology/Oncology, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
10Department of Anesthesia and Perioperative Care, University of California, San Francisco, School of Medicine, San Francisco, CA 94143, USA
11Division of Pulmonary, Critical Care, Allergy, and Sleep Medicine, University of California, San Francisco, School of Medicine, San Francisco, CA 94143, USA
12Cardiovascular Research Institute, Department of Medicine, University of California, San Francisco, School of Medicine, San Francisco, CA 94158, USA
13Division of Nephrology and Critical Care Medicine, University of California, San Francisco, School of Medicine, San Francisco, CA 94143, USA
14Department of Laboratory Medicine, University of California, San Francisco, School of Medicine, San Francisco, CA 94143, USA
15Division of Pulmonary Sciences and Critical Care Medicine, Department of Medicine, University of Colorado School of Medicine, Aurora, CO 80045, USA
16Division of Pulmonary, Allergy and Critical Care Medicine, Department of Medicine, University of Alabama at Birmingham, Birmingham, AL 35294, USA
17Pulmonary Section, Birmingham Veteran’s Affairs Medical Center, Birmingham, AL 35233, USA
18Section on Pulmonary and Critical Care, Allergy, and Immunology, Department of Internal Medicine, Wake Forest School of Medicine, Winston-Salem, NC 27157, USA
19Section on Hospital Medicine, Department of Internal Medicine, Wake Forest School of Medicine, Winston-Salem, NC 27157, USA
20Parker Institute for Cancer Immunotherapy, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA
21These authors contributed equally
22Lead contact
Corresponding authors: Allison R. Greenplate, E. John Wherry, Dokyoon Kim
Abstract
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.
논문정보
- Research article
- 2024년 11월 (BRIC 등록일 2024-11-18)
- 국외 연구진
관련 웨비나
![정확하고 해석가능한 세포 분석법 (Cytometry) 분석 인공지능 모델 개발 [Cell Rep. Med.]](/upload/board/files/onseminar/thumb/thumb_0000843.png)
학술웨비나
정확하고 해석가능한 세포 분석법 (Cytometry) 분석 인공지능 모델 개발 [Cell Rep. Med.] Flow or Mass Cytometry을 통해 얻어진 single-cell protein expression 데이터는 감염성 질환, 자가면역, 암, 그리고 치료 후 면역 세포 반응의 역할을 이해하는 데 필수적입니다. 특히 cytometry는 scRNA-seq과 비교해 수백만 개의 세포를 분석할 수 있어 드문 세포 집단을 식별하고 특성화하는 데 유리합니다. 본 발표에서는 single-cell cytometry 데이터를 분석하고 면역 프로파일링을 자동화하기 위해 인공지능 기반의 모델, cytometry masked autoencoder (cyMAE)를 소개합니다. cyMAE는 레이블이 없는 대규모 데이터셋에서 단백질 마커 간의 관계를 사전 학습(self-supervised pre-training)하여 추가적인 세포 정체성 정보 없이도 마커 간의 상호작용을 학습합니다. 이후, 사전 학습된 모델을 다양한 하위 과제에 맞게 미세 조정(fine-tuning)하여 독립적인 데이터셋에서도 복잡한 세포 유형을 정확히 식별하고, 단백질 마커의 중요성을 해석하며, 개인 수준의 정밀 면역 프로파일링을 가능하게 합니다.- 김재식(University of Pennsylvania)
댓글 작성 로그인
팝업 닫기관련 링크
연구자 키워드
연구자 ID
관련분야 연구자보기
소속기관 논문보기
관련분야 논문보기
해당논문 저자보기
