The Research On The Screening Of Pathogen- Ic Mechanism Of Glioblastoma Multiforme

Glioblastoma multiforme is the most common and aggressive malignant primary tumor of the adult central nervous system in the world,accounting for more than 50% of all gliomas.Although current treatments have continued to advance,they have not yet achieved satisfactory results in clinical practice.The prognosis of GBM patients is poor,with a median survival of approximately 12 months.In order to improve the prognosis of GBM patients,oncologists have conducted a lot of research,but so far,its etiology,pathogenesis and prognosis are still not fully understood.A large number of studies have shown that glioblastoma multiforme is an extremely complex biological phenomenon that is eventually formed through multiple stages.In these stages,it will be influenced by genetic factors and protein factors.However,it is still unclear which genes and proteins have a certain influence on the survival time of patients with GBM.Therefore,it has great significance to explore the key genes and proteins of GBM for improving the prognosis and targeted therapy.This paper constructs a multi-scale research model for GBM pathogenic mechanism,to find out the key genes and proteins that have important influence on GBM.(1)Screening GBM key genes.This section is mainly divided into 3 steps:(a)Data collection and preprocessing;(b)Obtaining differential genes;For the preprocessed data,CoxLasso algorithm,CoxSis algorithm and CoxSisLasso algorithm are used for survival analysis to find differential genes that have significant influence on the survival time of GBM patients.(c)Obtaining key genes;The intersection of different genes obtained by the three algorithms are taken as the final GBM key genes.(2)RPPA experimental design and data preprocessing.This section is mainly divided into 2 steps:(a)According to the principle of experimental design,RPPA experiments use reversed-phase protein array technology for two cell lines U251 and LN229 by changing the expression of AEBP1 gene to observe the protein expression,and thus obtain RPPA data;(b)Integrating three data preprocessing methods.This article integrates three data preprocessing methods,namely NormLinear,NormLog2 and NormLog2_MedianCentered.The data distribution is often skewed and the standard deviation is large.Some random errors occur in the analysis results of different data preprocessing methods.Therefore,this article takes the intersection of three results as the final result,which will greatly improve the accuracy and credibility of the results.(3)Screening GBM key proteins.This part is mainly based on the key gene AEBP1 screened in the first step,deducing the key proteins associated with the AEBP1 gene,which is the key proteins that affect GBM.It is mainly divided into 2 steps:(a)Obtaining proteins with significant differences.Due to the long cycle and high cost of biological experiments,the sample size of RPPA data in this paper is too small.Therefore,using a common statistical test to obtain differential proteins will cause the result to have certain deviation.This paper uses a combination of permutaion test and block design to solve the problem of small sample size,so as to obtain significantly different proteins.(b)Key proteins.For the purpose of GBM targeted pharmaceuticals,the differential proteins obtained are further analyzed for survival.Here,we use separately Cox regression algorithm,Conditional Quantile Screening algorithm,and Cbioportal website to analyze.Then,proteins,which are obtained by the intersection of three results,are taken as the key proteins of GBM.In this paper,a multi-scale study model is constructed by using real biological experimental data to screen key genes and key proteins of GBM.In conclusion,it can inprove the prognosis of patients by increasing the understanding of GBM pathogenic mechanisms.This not only lays the foundation for GBM targeted therapy,but also provides new ideas for further theoretical research.