单样本网络熵与某些复杂疾病恶性突变的预警

Single-Sample Network Entropy and Early Warning of Sudden Deterioration for Some Complex Diseases

摘要: 关于复杂疾病的早期诊断，传统方法需要许多病人样本来获取统计指标。而现实生活中人们不会频繁出入医院进行体检，故难以获得大量的患者数据，只有对个体单次取样的结果，这就需要提出基于单样本的探测复杂疾病发展过程中临界点的算法。本文基于动力系统和动态网络生物标志物的理论，利用个体新样本结合局部生物分子网络以及熵的概念开发了一个新算法。将该算法应用于数值模拟数据和一个真实的疾病数据，在这些应用中，该算法能及时准确地探测出临界点。本文开发的新算法具有稳定性和高效性，能帮助医学工作者进一步了解复杂疾病的动态生物分子机制，能更快速地提出合理的治疗方案。

Abstract: For the early diagnosis of complex diseases, traditional methods require many patient samples to obtain statistical indicators. In real life, people do not go to hospitals for medical examination frequently, so it is difficult to obtain a large number of patients data, and only the result of indi-vidual single sampling has been got, which requires a single sample based algorithm to detect critical points in the development of complex diseases. Based on the theory of dynamic system and dynamic network biomarkers, a new algorithm is developed by using new individual samples combined with local biomolecular networks and the concept of entropy. The algorithm is applied to numerical simulation data and a real disease data. In these applications, the algorithm can detect the critical point in time and accurately. The new algorithm developed in this paper is stable and efficient. It can help medical workers to further understand the dynamic biomolecular mechanism of complex diseases, and can more quickly propose a reasonable treatment plan.