Researchers from the Hebrew University of Jerusalem have made a groundbreaking advancement in the field of disease classification and treatment strategies by developing a machine learning model to identify disease subtypes. Led by Ph.D. student Dan Ofer and Prof. Michal Linial from the Department of Biological Chemistry at The Life Science Institute at Hebrew University, the study marks a significant progress in utilizing artificial intelligence in medical research.
The research, published in the Journal of Biomedical Informatics, highlights the importance of distinguishing diseases into distinct subtypes for accurate study and effective treatment strategies. While the Open Targets Platform integrates various datasets to support disease classifications, many disease annotations remain incomplete, especially for rare and orphan diseases.
The team introduced a novel machine learning approach to identify diseases with potential subtypes. Leveraging a database of approximately 23,000 diseases from the Open Targets Platform, they developed new features to predict diseases with subtypes using direct evidence. Machine learning models were then utilized to analyze feature importance and evaluate predictive performance, uncovering both known and novel disease subtypes.
The model demonstrated an impressive 89.4% ROC Area Under the Receiver Operating Characteristic Curve in identifying known disease subtypes. By integrating pre-trained deep-learning language models, the model’s performance was further enhanced. The research identified 515 disease candidates predicted to have previously unannotated subtypes, offering new insights into disease classification.
According to Ofer, “This project showcases the potential of machine learning in expanding our understanding of complex diseases. Advanced models can reveal hidden patterns and subtypes, leading to more precise and personalized treatments.” The innovative methodology provides a robust and scalable approach for improving knowledge-based annotations and offers a comprehensive assessment of disease ontology tiers.
The team is enthusiastic about the potential of their machine learning approach to revolutionize disease classification and ultimately enhance treatment strategies. This development signifies a significant step forward in utilizing artificial intelligence to advance medical research and improve patient outcomes.
