Characteristic Gene Mining For Tumor Cell Trajectories

With the development of single-cell RNA sequencing(Sc RNA-seq)technology,single-cell sequencing has been widely used to detect differential genes,and it can detect key characteristic genes in the process of tumor progression.Because most brain tissue samples are frozen,and it is difficult to obtain fresh brain tissue for study.Sc RNA-seq is very suitable for brain tumor analysis.The objective of this study was to screen the characteristic genes of glioblastoma(GBM)based on trajectory analysis and survival analysis of single cell RNA sequencing.In this study,18 cases of glioblastoma single-cell sequencing data from the Gene Expression Omnibus(GEO)data base were filtered by quality control and principal component analysis.The filtered genes were reduced in dimension,and the differential expression genes and functional enrichment analysis were screened by clustering and trajectory analysis;Immunological scoring was performed using 167 GBM transcripts from The Cancer Genome Atlas(TCGA)data base,divided into two groups by median value,and differential analysis was performed.The intersection of differential genes analyzed in GEO and TCGA data base was selected,and the expression values of these genes in TCGA data base were used for univariate Cox risk analysis combined with 593 GBM clinical data downloaded from TCGA data base.Then,significant genes were screened for risk scores by LASSO regression and multivariate Cox risk analysis.The TCGA transcription data were divided into high and low risk groups according to the median risk score.Finally,the characteristic genes of GBM were determined by survival analysis combined with significant genes.The characteristic genes screened in this study play an important role in tumor proliferation,metastasis and occurrence through data review.Some genes have not been studied in GBM,but they play a key role in other cancers,and the survival probability of high-risk patients decreases rapidly over time when analyzed with clinical data.