Bioinformatic methods uncover 5 diagnostic biomarkers associated with drug resistance and metastasis for gastrointestinal stromal tumor

Gastrointestinal stromal tumor (GIST) is a prevalent neoplasm in digestive tract, drug resistance and metastasis remain the major challenges during treatment. This work commenced to identify novel diagnostic biomarkers associated with drug resistance and metastasis in GIST using bioinformatic analysis. Single-cell and RNA-sequencing data of GIST were collected from Gene Expression Omnibus (GEO) database. Single cell atlas of GIST was analyzed employing “Seurat” package. Differentially expressed genes (DEGs) were screened by FindMarkers function and “limma” package, functional enrichment analysis was performed with “clusterProfiler” package. Hereafter, drug resistance-related module genes were identified using “WGCNA” package and metastasis-associated DEGs were recognized by protein-protein interaction (PPI) network analysis. Further, the biomarkers of GIST were identified via machine learning, the diagnostic values of which were assessed by receiver operating characteristic (ROC) curve. Finally, the correlation between biomarkers and immune infiltration was analyzed by Pearson method. This study revealed 8 cell types of GIST, the proportion of Myeloid and natural killer (NK)/T cells were elevated in metastatic/resistance GIST samples. Further, we screened 1074 drug resistance-related module genes and 667 metastasis-related DEGs. Drug resistance-related DEGs were mainly involved in immune and inflammatory response pathways, while metastasis-related DEGs were primarily implicated in xenobiotics and drug metabolism pathways. Thereafter, 5 biomarkers ( ADAMTS1 , BATF3 , APOC1 , AP1S3 , SLC22A18AS ) were identified, exhibiting good diagnostic potential in GIST. These biomarkers showed positive correlation with most immune cell infiltration. This study uncovered 5 diagnostic biomarkers related to drug resistance and metastasis in GIST, which may help to improve therapeutic strategies for GIST.