Morphing projections: a new visual technique for fast and interactive large-scale analysis of biomedical datasets

Area of Science:

• Biomedical data analytics
• Bioinformatics
• Computational biology

Background:

• Biomedical research involves analyzing high-dimensional data with clinical metadata.
• Current data analytics software offers limited user interaction and flexibility.
• Complex data analysis requires fluid user feedback for cognitive enhancement.

Purpose of the Study:

• To present a new technique for biomedical data analytics.
• To enable efficient blending of meaningful data views for interactive exploration.
• To facilitate the discovery of relationships and hypothesis generation in high-dimensional biomedical data.

Main Methods:

• Developed a technique blending diverse data representations.
• Implemented an interactive interface for smooth transitions between data views.
• Utilized gene expression data and clinical metadata in case studies.

Main Results:

• Demonstrated a natural and smooth transition among complementary data representations.
• Showcased the potential for discovering relevant relationships through interactive exploration.
• Successfully generated new hypotheses from high-dimensional, multidomain biomedical data.

Conclusions:

• The proposed technique enhances the cognitive process of biomedical data analysis.
• Interactive blending of diverse data views is key to discovering novel insights.
• The approach is effective for analyzing complex datasets like gene expression and clinical metadata.