A Multi-Dimensional Approach to Map Disease Relationships Challenges Classical Disease Views.

Area of Science:

• Computational biology
• Bioinformatics
• Systems medicine

Background:

• Traditional disease classification relies on organ systems and phenotypes, overlooking crucial mechanistic links for therapeutics.
• Understanding disease relatedness is vital for drug repurposing and discovering new therapeutic strategies.

Purpose of the Study:

• To develop a multi-view disease map integrating molecular, clinical, and pharmacological data to reveal novel disease relationships.
• To advance disease understanding beyond conventional classifications and identify shared mechanistic features.

Main Methods:

• Mapped similarity among 502 diseases using six data dimensions (molecular, clinical, pharmacological).
• Employed multiple distance measures and multi-view clustering for disease relatedness patterns.
• Integrated data dimensions into a consensus map for a comprehensive disease view.

Main Results:

• A multi-view disease map revealed novel relationships divergent from the International Classification of Diseases (ICD).
• Identified enriched disease features like genes, pathways, and chemicals within distinct disease groups.
• Found concordance with known associations and uncovered shared features between Type 2 diabetes and Alzheimer's disease.

Conclusions:

• The multi-view disease map provides novel insights into disease relatedness, challenging traditional views.
• This approach facilitates the reconstruction of comorbidity patterns and aids in drug repurposing.
• Revising disease relationships holds promise for advancing drug discovery and personalized medicine.