Dynamic Network Biomarkers And Early Warning Of Diseases

In many complex diseases,the transition from a healthy state to a diseased state does not happen suddenly,but a process of dynamic change.Acute lung injury is a common clinical health problem,it threatens the lives of patients with high morbidity and mortality.Therefore,identifying the key points of disease deterioration and biomarkers are essential for effective treatment.Generally,molecular biomarkers of diseases are obtained based on differences in molecular expression levels to distinguish the normal state of the disease from the disease state.Because the dynamic network biomarker method can detect dynamic mutations in the molecular characteristics of complex diseases,it can monitor and evaluate different stages of the disease.This study explores the critical state of disease deterioration and dynamic network biomarkers by constructing early warning signals to identify the key biomarkers of the disease process.The main contents are as follows:In the second chapter,in order to identify the critical state of acute lung injury worsening and the key genes in the differentially expressed genes that affect the occurrence and development of the disease,a single-sample dynamic network biomarker discovery method based on the K-means clustering algorithm is proposed.Firstly,a single sample-specific network is constructed based on the difference in Pearson correlation coefficient between normal samples and new cases.Secondly,the K-means clustering algorithm is used to decompose the network into multiple modules,and the early warning signal is constructed according to the dynamic network biomarker method,and the score of each module is calculated.Finally,the criticality and biomarkers of the disease are determined according to the maximum score of the module at each moment.The experiment found that the critical state was at 8 h,and gene function analysis showed that the obtained biomarkers were related to cell senescence,apoptosis,and inflammation.Use the MCC algorithm to select the top 10 key genes with the largest cluster centrality for analysis,found that they play a positive regulatory role in the disease process,and are related to cell proliferation,stress response,cancer progression,and pulmonary fibrosis,further verifying the effectiveness of this method.In the third chapter,considering that dynamic network biomarkers may or may not be differentially expressed genes,the top 500 highly biased genes are selected for analysis based on the differentially expressed information of case samples and normal samples.Firstly,given the initial solution and several candidate solutions are obtained through the neighborhood solution generation method.Secondly,the initial solution and candidate solution scores are calculated by the early warning signal which is constructed according to the dynamic network biomarker method.The current solution is updated through the tabu search algorithm until the end of the iteration.Finally,the critical state and biomarkers of acute lung injury are determined through the optimal solution of the algorithm.The experiment found that the critical state was at 8 h,which is consistent with the results in chapter 2.Furthermore,gene function analysis of biomarkers before and after the critical point,PPI network analysis and literature verification showed that the appearance of inflammatory response,immune response and immune disorders may mean that the disease state is about to change.