dendsort: modular leaf ordering methods for dendrogram representations in R.

Area of Science:

• Bioinformatics
• Computational Biology
• Data Visualization

Background:

• Dendrograms visualize hierarchical clustering, crucial for life science data analysis.
• Cluster heat maps use dendrogram leaf order for data reordering, but current methods have limitations.
• Existing leaf ordering can obscure relationships and offers multiple ambiguous arrangements.

Purpose of the Study:

• To develop novel modular leaf ordering methods for dendrograms.
• To more faithfully encode monotonic merge order and nested structures.
• To improve the interpretability and legibility of dendrogram and heat map visualizations.

Main Methods:

• Proposed two modular leaf ordering heuristics.
• Implemented methods in an R package named "dendsort".
• Compared visualizations against default R heuristics and seriation methods.

Main Results:

• The new methods yield dendrogram structures with clearer global patterns.
• Visualizations are more interpretable, especially in limited display spaces.
• The proposed heuristics offer more insightful representations for certain datasets.

Conclusions:

• The developed leaf ordering techniques enhance dendrogram and heat map utility.
• These methods improve the visualization of hierarchical relationships in biological data.
• The "dendsort" R package provides accessible tools for applying these advancements.