隐私保护下单细胞RNA测序数据细胞分类研究

Study on Cell Classification of Single-cell RNA Sequencing Data under Privacy Protection

目的/意义研究安全的高维稀疏数据处理方法,提高分析精度并保障敏感信息安全,促进单细胞RNA测序技术的广泛应用。方法/过程提出基于可信执行环境(trusted execution environment,TEE)的解决方案,将训练数据加密后传输至TEE,在安全隔离环境中解密并训练,获得训练后的模型参数。对比分析TEE和传统明文环境下使用基于神经网络的自动细胞类型识别模型和支持向量机进行细胞分类的表现。结果/结论TEE下两种分类模型的F 1分别达到0.904和0.879,与传统明文环境下性能相当;TEE提供的安全执行环境对模型的准确性和效率影响极有限,可用于处理敏感或私有数据场景。

Purpose/Significance To develop a secure single-cell RNA sequencing(scRNA-seq)classification method,which can enhance data analysis precision and ensure the security of sensitive information,and to promote the application of scRNA-seq technology in various fields.Method/Process The paper proposes a solution based on trusted execution environment(TEE).The training data is encrypted and transmitted to TEE.It is decrypted in a secure and isolated environment,while training the model to obtain the trained model parameters.Automated cell type identification using neural networks(ACTINN)and support vector machine(SVM)are used for cell classification in both TEE and traditional plaintext environments.The results are compared and analyzed.Result/Conclusion The results show that the F 1 score of the two classification models in TEE environment reaches 0.904 and 0.879,respectively,which is comparable to the performance in traditional plaintext environment.The secure execution environment provided by TEE has extremely limited impact on the accuracy and efficiency of the models.This is of great significance for seeking both secure and efficient data processing solutions in scenarios where sensitive or private data needs to be processed.