Abstract:
Introduction: Inter-individual cancer variability remains the main challenge for resistance to drug treatment. One of the well-known reasons is the set of genetic alterations and polymorphisms affecting target genes, which cause subsequent changes in gene expression patterns among patients. Whether the observed variability in gene expression affects diagnosis, prognosis, and outcome remain poorly understood. For this, understanding the molecular mechanisms underlying biological variability entails identifying the set of variable and non-variable genes in different cancer types.
Aim: In this study, we hypothesize that biological variability metric will identify key genes that play a role in cancer diagnosis, progression and drug-response.
Methods: Biological variability was calculated on patients' transcriptome data across 33 different cancers retrieved from the TCGA database. Profiling the transcriptome in individuals using RNA-seq technologies has been widely used to obtain mRNA-based molecular markers. Given the robustness of RNA-seq data, we propose a metric that can easily be implemented to detect molecular biomarkers, whether diagnostic, prognostic, or therapeutic using gene expression data...
Results: We derived a list of prognostic and diagnostic markers that were cancer type-specific or common between cancers. We then derived all the list of potential drug-target genes based on their biological variability score and oncogenic properties
Conclusion: Not only is biological variability an important measure to identify variable and non-variable genes in each cancer, but it is key to identify cancer-specific molecular markers, predict reliable drug-target genes and identify genes related to cancer progression and development. Gaining a deeper understanding of genes expression variability in cancer will broaden our knowledge on genes related to resistance early cancer detection, outcome, and the development of personalized treatment.