A data driven approach reveals disease similarity on a molecular level

Area of Science:

• Computational Biology
• Systems Biology
• Bioinformatics

Background:

• Understanding molecular similarities across diverse diseases is crucial for identifying common biological mechanisms.
• High-dimensional, low-sample 'omics' data present challenges for comparative analysis.

Purpose of the Study:

• To develop and apply a novel method for computing similarities between statistical distributions of 'omics' datasets.
• To visualize the biological data landscape and uncover non-trivial connections between different diseases and studies.

Main Methods:

• Developed a statistical method to compare empirical distributions of high-dimensional, low-sample datasets.
• Applied the method to hundreds of 'omics' studies, creating dataset-to-dataset and disease-to-disease networks.
• Implemented a technique to identify key molecular quantities and pathways driving observed similarities.

Main Results:

• Generated networks visualizing the landscape of biological data, revealing dataset interconnections.
• Discovered significant links between diseases such as Alzheimer's and schizophrenia, asthma and psoriasis, and liver cancer and obesity.
• Identified molecular drivers and pathways underlying these cross-disease similarities.

Conclusions:

• The findings highlight shared molecular mechanisms across seemingly unrelated diseases.
• The developed method acts as a 'statistical telescope' for exploring biological data.
• This approach can reveal novel drug targets and deepen our understanding of complex diseases.