发表机构
Institute for Bioinformatics and Medical Informatics, University of Tübingen; Medical Data Privacy and Privacy-Preserving Machine Learning, University of Tübingen; University Hospital Aachen; Delft University of Technology（蒂宾根大学生物信息学与医学信息学研究所; 蒂宾根大学医学数据隐私与隐私增强机器学习研究所; 亚琛大学医院; 代尔夫特理工大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究针对跨机构共享单细胞数据的隐私问题，提出基于安全多方计算的CellCnn训练推理框架，保留模型关键组件，在CMV和AML数据集上实现接近明文的准确率，优于现有隐私保护基线。
AI 中文摘要
从高维单细胞测量数据中检测罕见疾病相关细胞亚群，对于理解白血病、病毒感染等疾病至关重要。CellCnn是专为该任务设计的卷积神经网络（CNN），已证明能识别频率低至0.01%的表型相关细胞群体。可靠训练此类模型需要比单个机构通常能汇集的更大、更多样化的患者队列，且底层单细胞数据过于敏感，无法在现有隐私法规下跨机构共享。我们提出一种安全多方计算（MPC）框架，可在秘密共享数据上完全实现CellCnn的训练与推理，确保参与者和计算服务器均无法观察到原始患者数据或中间值。在巨细胞病毒感染（CMV）和急性髓系白血病（AML）的基准单细胞数据集上评估，我们的实现保留了与明文对应模型相近的准确率，且优于现有隐私保护基线。与早期移除ReLU激活、偏置项等组件的隐私保护方法不同，我们的方法保留了CellCnn架构的这些关键部分，在不暴露原始患者数据的情况下支持准确分析。
英文摘要
The detection of rare disease-associated cell subsets from high-dimensional single-cell measurements is critical for understanding diseases such as leukaemia and viral infections. CellCnn, a convolutional neural network (CNN) designed for this task, has demonstrated the ability to identify phenotype-associated cell populations at frequencies as low as 0.01\%. Training such models reliably requires patient cohorts that are larger and more diverse than any single institution can typically assemble, and the underlying single-cell data is too sensitive to share across institutional boundaries under existing privacy regulations. We propose a secure multi-party computation (MPC) framework that enables the training and inference of CellCnn entirely on secret-shared data. This ensures that neither the participants nor the computing servers ever observe raw patient data or intermediate values. Evaluated on benchmark single-cell datasets for cytomegalovirus infection (CMV) and acute myeloid leukaemia (AML), our implementation preserves accuracy close to its plaintext counterpart while outperforming the prior privacy-preserving baseline. In contrast to earlier privacy-preserving approaches that removed components such as ReLU activations and bias terms, our method retains these key parts of the CellCnn architecture and supports accurate analysis without exposing raw patient data.
CommentsAccepted at the CIBB 2026 conference (https://cibb2026.teralab.ai/)
