LeMeDISCO

LeMeDISCO predicts disease comorbidities by identifying shared mode-of-action proteins to provide molecular interpretations of disease co-occurrence.

Key Features:

• AI-based prediction: Uses the MEDICASCY algorithm to identify proteins with shared modes of action across diseases and predict comorbidities at large scale.
• Comprehensive disease analysis: Applied to predict comorbidity occurrences among 3,608 distinct diseases.
• Benchmarking performance: Reports comorbidity recall rates of 44.5% versus 6.4% for the XD-score, 68.6% versus 8.0% for the S_AB score, and 63.7% versus 100% versus the Symptom Similarity Score.
• Molecular mechanism insights: Focuses on shared proteins to identify essential proteins and pathways underlying disease comorbidity.

Scientific Applications:

• Research domains: Applicable to genomics, proteomics, and systems biology investigations of disease relationships.
• Therapeutic target identification: Pinpoints proteins that are potential targets across multiple co-occurring diseases.
• Pathway and mechanism analysis: Enables elucidation of biological pathways implicated in disease comorbidity and multi-disease mechanisms.
• Predictive modeling: Augments models of disease progression and patient outcomes with molecular comorbidity information.

Methodology:

Integrates artificial intelligence via the MEDICASCY algorithm to identify shared mode-of-action proteins and predict comorbidities, applied to 3,608 diseases, and benchmarked against the XD-score, S_AB score, and the Symptom Similarity Score using reported comorbidity recall rates.