Predict effective drug combination by deep belief network and ontology fingerprints

Area of Science:

• Computational biology
• Bioinformatics
• Artificial intelligence in medicine

Background:

• Synergistic drug combinations are highly desirable for cancer treatment.
• Systematically predicting effective drug combinations remains a significant challenge in oncology.

Purpose of the Study:

• To develop and validate a novel computational method for predicting drug synergy.
• To integrate multi-omics data with literature-derived gene profiles for enhanced prediction accuracy.

Main Methods:

• Utilized a deep belief network (DBN) model.
• Integrated gene expression, pathway information, and Ontology Fingerprints (literature-derived gene profiles).
• Applied the method to drug synergy data from the 2015 DREAM competition.

Main Results:

• The integrative DBN method demonstrated superior performance in predicting drug synergy compared to published results.
• Successfully predicted synergy for 4999 drug pairs, validating the approach.
• Showcased the feasibility of combining -omics data with literature insights for drug synergy prediction.

Conclusions:

• The developed artificial intelligence approach effectively predicts drug synergy using integrated biological data.
• This method offers a promising strategy for identifying effective cancer drug combinations.
• Highlights the potential of advanced computational techniques in drug discovery and personalized medicine.