面向多组学癌症数据的多视图生存分析算法的研究

癌症是威胁人类生命健康的重要因素，探索癌症的产生和发展机制有利于对癌症进行有效的治疗。生存分析算法可以有效地探索癌症患者的生存时间与各种不同影响因素之间的内在关系，进而为癌症的研究提供更有价值的参考信息。同时，多组学数据从不同的视角描述了疾病的不同特性，也为癌症的研究和相关算法的建模提供了更加全面的数据支持。基于此，本项目将面向多组学癌症数据，结合多视图学习算法机制，提出多视图生存分析算法框架。通过兼容不同组学数据之间的各自优势和规避各自局限、探索癌症亚型和挖掘生物标志物等内容的研究，解决多组学癌症数据所关联的关键科学问题，为癌症的临床研究建立更加准确的生存分析模型并有效地挖掘生存时间与各种因素之间的内在联系，从而实现能够为更加深刻地认识疾病提供有效的算法支持的研究目标。

Cancer has a significant impact on public health. Discovering the progression of disease is beneficial to provide more precise treatment for patients. Survival analysis is an effective way to exploit the underlying relationship between survival time and clinical features and thus is helpful to the understanding of disease progression. In addition, advances of high throughput experimental methods have led to the availability of more diverse omic datasets in clinical analysis application. It is important for clinical analysis to exploit different types of omic data simultaneously. Toward this end, this proposal aims to propose a multi-view survival analysis algorithm framework for multi-omic cancer data application by exploring consistent and complementary information among multi-omic data, identifying meaningful cancer subtypes and biomarker information. As a result, the proposed multi-view survival analysis framework is able to identify biologically meaningful information for clinical research.