The cancer genome explorer (TCGEx): a powerful visual interface for sophisticated analyses of high throughput cancer data

Abstract

Analyzing high-throughput genomics data requires programming expertise, and it remains challenging for many experimental researchers. While visual interfaces have eased access to this data, their limited flexibility in accommodating complex custom analyses remains a significant hurdle. To address these shortcomings, we have developed The Cancer Genome Explorer (TCGEx), a web-based R/Shiny application that can work with preprocessed The Cancer Genome Atlas (TCGA) transcriptomics data and user-provided external datasets. TCGEx serves as a centralized hub, offering a diverse array of analytical tools including survival modeling, exploratory graphing, gene set enrichment, gene-to-gene correlation, dimensionality reduction, and machine learning. Our utilization of TCGEx in investigating gene expression profiles within human primary and metastatic melanoma revealed distinctive tumor subsets characterized by unique immune signatures and survival outcomes. Delving deeper, we explored miRNA networks associated with intratumoral immunity, harnessing TCGEx's machine learning algorithms. Aligning with existing literature, our study highlighted miR-155 as among the prominently upregulated miRNAs in immune-enriched melanoma biopsies. Intriguingly, heightened miR-155 levels correlated with transcriptomic enrichment in lipid catabolism pathways and depletion in ribonucleoside catabolism pathways. Expanding our inquiry to previously published datasets from melanoma immunotherapy trials, we discerned transcriptomic patterns linked to therapeutic benefits. While our study predominantly focused on immune-associated noncoding RNAs within the melanoma tumor microenvironment, TCGEx extends its capabilities to investigate 32 other TCGA cancer projects as well as user-uploaded external datasets. In essence, TCGEx emerges as a powerful and adaptable platform facilitating the analysis of high-throughput cancer data.