Cat-E: A comprehensive web tool for exploring cancer targeting strategies

Abstract

Identifying potential cancer-associated genes and drug targets from omics data is challenging due to its diverse sources and analyses, requiring advanced skills and large amounts of time. To facilitate such analysis, we developed Cat-E (Cancer Target Explorer), a novel R/Shiny web tool designed for comprehensive analysis with evaluation according to cancer-related omics data. Cat-E is accessible at https://cat-e.bioinfo-wuerz.eu/. Cat-E compiles information on oncolytic viruses, cell lines, gene markers, and clinical studies by integrating molecular datasets from key databases such as OvirusTB, TCGA, DrugBANK, and PubChem. Users can use all datasets and upload their data to perform multiple analyses, such as differential gene expression analysis, metabolic pathway exploration, metabolic flux analysis, GO and KEGG enrichment analysis, survival analysis, immune signature analysis, single nucleotide variation analysis, dynamic analysis of gene expression changes and gene regulatory network changes, and protein structure prediction. Cancer target evaluation by Cat-E is demonstrated here on lung adenocarcinoma (LUAD) datasets. By offering a user-friendly interface and detailed user manual, Cat-E eliminates the need for advanced computational expertise, making it accessible to experimental biologists, undergraduate and graduate students, and oncology clinicians. It serves as a valuable tool for investigating genetic variations across diverse cancer types, facilitating the identification of novel diagnostic markers and potential therapeutic targets.

Keywords: Cancer; Cancer pathways; Immune modulation; Network analysis; Oncolytic virus; Protein interaction.

Graphical abstract

Fig. 1

Schematic representation of Cat-E web tool analyses workflow. Cat-E implements nine functional analyses related to cancer and oncolytic viruses. Blue arrows depict the logical flow of an analysis.

Fig. 2

Application of Cat-E for cancer research. Summary of analytical parameters and outputs for the LUAD case studies performed using Cat-E. Potential targets for anti-cancer drug action are explored using the databases in Cat-E.

Fig. 3

A. Volcano plot illustrating DGEs in LUAD (the black, blue, and red circles represent nonsignificant, down-regulated, and up-regulated genes, respectively); B. A Venn diagram showing the comparison between the LUAD data within Cat-E (Selected_Data) and the DGE LUAD data in the GEPIA2 (User_Data); C. A bar graph displaying the top 20 GO terms obtained from enrichment analysis; D. Visual representation of mutation positions within the muc16 gene in LUAD.

Fig. 4

A. Survival analysis based on pathological staging in LUAD clinical data, with each color representing a distinct pathological stage (Red: Stage1, Blue: Stage1A, Green: Stage1B, Purple: Stage2, Orange: Stage2A, Yellow: Stage2B, Brown: Stage3A, Pink: Stage3B, Grey: Stage4); B. Comprehensive analysis of overall survival regarding the expression levels (Green: down-regulation, Orange: up-regulation) of muc16 in LUAD tumor samples; C. Identification of active metabolic pathways within LUAD data, illustrated by a red line indicating log₂FC> 0 and a green line indicating log₂FC< 0.

Fig. 5

Analysis of the metabolic flux in LUAD cancer cells. A. Module for loading or generating the necessary.sbml file essential for conducting metabolic flux analysis; B. Representation of the generated metabolic model (additional tabs furnish results of flux analysis, optimized outcomes, and summary data); C. Computed flux values for the reactions generated, presented as a network.

Fig. 6

Interactive model of the pck2 (Q16822) protein in Cat-E, generated using AlphaFold for protein structure prediction. The visualization highlights the first 20 amino acids of the protein sequence (selection: 1–20), depicted in yellow. The color scheme applied to the protein structure is based on AlphaFold confidence scores, with dark blue representing very high confidence (pLDDT > 90), light blue indicating high confidence (90 > pLDDT > 70), yellow reflecting low confidence (70 > pLDDT > 50), and orange denoting very low confidence (pLDDT < 50). This confidence-based coloring scheme facilitates an in-depth assessment of structural predictions.