EZ-Pair graph: scalable unified-axis visualization method for summarizing large-scale paired data

Area of Science:

• Bioinformatics
• Data Visualization
• Computational Biology

Background:

• Dual-axis visualizations hinder quantitative interpretation of large-scale paired datasets.
• Conventional single-axis plots risk visual saturation with numerous paired lines.
• Existing methods struggle to effectively display summary metrics with raw data.

Purpose of the Study:

• To develop scalable visualization methods for large-scale paired data.
• To improve the interpretability of complex datasets by integrating summary metrics.
• To address limitations of dual-axis and single-axis plots in paired data analysis.

Main Methods:

• Developed EZ-Pair Graph, a suite of scalable methods for aggregating positional and slope information.
• Implemented three complementary tools: trapezoid plot, clustered line plot, and parallel arrow plot.
• Focused on summarizing paired differences, their rank, magnitude, and directional heterogeneity.

Main Results:

• EZ-Pair Graph effectively visualizes large-scale paired biological datasets.
• Revealed structured, localized, and heterogeneous trends missed by conventional methods.
• Demonstrated improved interpretability and detection of underlying patterns.

Conclusions:

• EZ-Pair Graph enhances the interpretation of distributional differences in large paired datasets.
• The methods are valuable for detecting subtle patterns and variations in complex biological data.
• Scalable visualization is crucial for advancing analysis of increasingly large datasets.