ECM yönlendirmeli entegre ağ modellemesiyle hasta siniflandirmasi

The extracellular matrix (ECM) plays a crucial role in tumor initiation, progression, and drug response. Consequently, the gene/protein expression signatures of ECM in tumors could serve as significant prognostic factors. This study aims to stratify lung adenocarcinoma (LUAD) patients using an ECM-guided multi-omic approach and further define ECM profiles of these groups through network modeling, providing deeper insights into prognosis and facilitating the selection of patient-specific therapies. Publicly available multi-omic datasets from 101 LUAD patients with paired tumor-normal adjacent tissue samples were used in this study. Tumors were labeled based on the multi-omic expression scores of ECM categories, followed by consensus clustering, resulting in four distinct patient clusters. For each patient, important ECM genes and transcriptional regulators were identified and utilized in network modeling. Various enrichment analyses were performed on the resulting networks, and scores were calculated for all enriched pathways, hallmarks, and transcriptional regulators for each patient. For each cluster, a consensus network was constructed using the patient networks, and a proximity analysis was performed to identify effective drug molecules. This study revealed patient groups representing different ECM grades, and these ECM grades showed distinct clinical features, mutation profiles, and cellular heterogeneity. It was shown that pathways closely relevant to metastasis like epithelial-to-mesenchymal transition and angiogenesis, cancer stemness like Wnt and hedgehog signaling, and cancer-related inflammation like TNF-alpha and NF-KB were also highly enriched in ECM grades of high severity. Drug screening revealed drugs that could be more effective on patients of a particular ECM grade. Overall, with this thesis, we were able to capture the heterogeneous nature of the ECM observed in tumors by partitioning patients with discrete ECM characteristics into groups that can be targeted with unique therapeutic approaches.