Background Cancer is gradually becoming a major global public health challenge, not only due to its widespread impact and increasing incidence, but also because each tumor type has its unique biological characteristics and complex pathogenesis. In recent years, advancements in bioinformatics have enabled researchers to study the molecular mechanisms of cancer from multiple dimensions, including genomics, transcriptomics, proteomics, and metabolomics. This study aims to explore effective bioinformatics methods for integrating multi-omics data and their applications. By investigating Mendelian Randomization (MR) techniques and lncRNA prognostic models, the study seeks to develop new approaches for cancer risk prediction, early detection, and prognostic assessment, in order to reveal the molecular regulatory mechanisms of cancer and identify potential therapeutic targets. Methods This study first used data from The Cancer Genome Atlas (TCGA) database to construct a prognostic model based on disulfidptosis-related lncRNAs. Next, MR was employed to systematically assess causal relationships between immune cells, blood metabolites, and cancer risk, focusing on lung and breast cancer. Additionally, a summary-based MR approach was used to explore the causal links between antihypertensive drug target genes and cancer risk. This method was further applied to a larger drug target gene database to identify cancer-related genes and potential opportunities for drug repurposing. Results First, the prognostic model based on disulfidptosis-related lncRNAs, combined with a large set of clinical data (including sex, age, Clinical Tumor stage (T stage), and other factors), significantly improved the accuracy of survival risk prediction in cancer patients. Additionally, the role of disulfidptosis-related lncRNAs in metabolic regulation and immune evasion was explored. Furthermore, significant associations were identified between specific immune cells, metabolites, and cancer outcomes, providing new insights into the immune and metabolic regulatory mechanisms underlying cancer development. Next, summary-based MR analysis revealed significant causal relationships between certain antihypertensive drug target genes and cancer risk. Finally, by exploring a more comprehensive drug target gene database, we identified additional cancer-related genes and uncovered opportunities for repurposing existing drugs for cancer treatment. Conclusion By integrating multi-omics data and combining lncRNA prognostic models with MR techniques, this study systematically elucidated the molecular mechanisms of cancer and provided new methodological insights for cancer risk prediction, early diagnosis, and personalized treatment. The lncRNAs, specific metabolites, immune cells, and drug target genes identified in this study offer potential directions for future cancer chemoprevention and the development of therapeutic targets.
Multi-Omics and Mendelian Randomization Studies Unveiling Genetic and Molecular Mechanisms of Cancer Risk and Prognosis
Huang, Y. (Author). 8 Jan 2025
Student thesis: Phd