CombiROC: an interactive web tool for selecting accurate marker combinations of omics data

Area of Science:

• Biotechnology
• Bioinformatics
• Computational Biology

Background:

• Diagnostic accuracy is crucial for disease identification.
• Combining multiple markers enhances diagnostic performance.
• Selecting optimal marker combinations is complex, requiring advanced statistical methods.

Purpose of the Study:

• To develop a user-friendly tool, CombiROC, for selecting optimal marker combinations from diverse omics data.
• To improve the accuracy of disease diagnosis through sophisticated marker signature analysis.
• To provide researchers with a flexible platform for analyzing omics data.

Main Methods:

• Developed CombiROC, a software tool with a graphic interface.
• Integrated sensitivity/specificity filters for analyzing data from proteomics and transcriptomics.
• Enabled user control over initial selection stringency and bypassed limitations of experimental approaches.
• Computed sensitivity, specificity, and ROC curves for all marker combinations.

Main Results:

• CombiROC optimizes the number of candidate marker panels from combinatorial analysis.
• The tool reduces computational burden and false negative rates by avoiding hard-coded thresholds.
• Validated with published data, confirming known marker combinations and discovering new ones.
• Provided automatic comparisons of best combination performances via ROC curves.

Conclusions:

• CombiROC is a novel, freely available tool for the scientific community.
• It simplifies the complex process of identifying optimal multimarker signatures.
• The tool enhances diagnostic accuracy and offers a flexible approach to omics data analysis.