RESUMEN
Rectal cancer (RC) presents significant treatment challenges, particularly in the context of chemotherapy resistance. Addressing this, our study pioneers the use of matched RC tumor tissue and patient-derived organoid (PDO) models coupled with the innovative computational tool, Moonlight, to explore the gene expression landscape of RC tumors and their response to chemotherapy. We analyzed 18 tissue samples and 32 matched PDOs, ensuring a high-fidelity representation of the tumor bioloy. Our comprehensive integration strategy involved differential expression analyses (DEAs) and gene regulatory network (GRN) analyses, facilitating the identification of 5,199 genes governing at least one regulon. By using the biological processes (BPs) collected from Moonlight closely related to cancer, we pinpointed 2,118 regulator-regulon groups with potential roles in oncogenic processes. Further, through integration of Moonlight and DEA results identified 334 regulator-regulon groups significantly enriched in both tissue and PDO samples, classifying them as oncogenic mediators (OMs). Among these, four genes (NCKAP1L, LAX1, RAD51AP1, and NAT2) demonstrated an association with drug responsiveness and recurrence-free survival (RFS), offering new insights into the molecular mechanisms of chemotherapy response in RC. Our integrated approach not only underscores the translational fidelity of PDOs, but also harnesses the analytical prowess of Moonlight, setting a new benchmark for targeted therapy research in rectal cancer.
RESUMEN
Colorectal cancer (CRC) poses significant challenges in chemotherapy response prediction due to its molecular heterogeneity. This study introduces an innovative methodology that leverages gene expression data generated from matched colorectal tumor and organoid samples to enhance prediction accuracy. By applying Consensus Weighted Gene Co-expression Network Analysis (WGCNA) across multiple datasets, we identify critical gene modules and hub genes that correlate with patient responses, particularly to 5-fluorouracil (5-FU). This integrative approach advances precision medicine by refining chemotherapy regimen selection based on individual tumor profiles. Our predictive model demonstrates superior accuracy over traditional methods on independent datasets, illustrating significant potential in addressing the complexities of high-dimensional genomic data for cancer biomarker research.