Discovering comorbid diseases using an inter-disease interactivity network based on biobank-scale PheWAS data.

Area of Science:

• Genetics and Bioinformatics
• Computational Biology
• Medical Informatics

Background:

• Understanding disease comorbidity is crucial for effective prevention, treatment, and prognosis.
• Identifying likely or unlikely disease co-occurrences can elucidate complex disease relationships.

Purpose of the Study:

• To introduce and utilize an inter-disease interactivity network for discovering and prioritizing comorbidities.
• To develop a comorbidity scoring algorithm for predicting disease co-occurrence based on directional genetic effects.

Main Methods:

• Constructed an inter-disease interactivity network to analyze disease associations.
• Determined disease associations by considering the direction of effects of shared genetic components.
• Categorized associations as synergistic or antagonistic and developed a comorbidity scoring algorithm.

Main Results:

• Investigated inter-disease associations among 427 phenotypes using UK Biobank PheWAS data.
• Predicted comorbidity priorities and verified findings using UK Biobank inpatient electronic health records.
• Demonstrated that considering phenotype association interactions improves comorbidity prediction.

Conclusions:

• The inter-disease interactivity network is a valuable tool for discovering and prioritizing comorbidities.
• The developed comorbidity scoring algorithm effectively predicts disease co-occurrence.
• Incorporating interaction effects of phenotype associations enhances comorbidity prediction accuracy.