MDIPA: a microRNA-drug interaction prediction approach based on non-negative matrix factorization

Area of Science:

• Biochemistry and Molecular Biology
• Computational Biology
• Pharmacology

Background:

• MicroRNAs (miRNAs) are small biomolecules regulating gene expression, crucial in disease development and treatment.
• Drug interactions with miRNAs can alter their functions, impacting therapeutic outcomes.
• Experimental identification of miRNA-drug interactions is costly and time-consuming, necessitating computational prediction methods.

Purpose of the Study:

• To propose an effective computational approach for predicting microRNA-drug interactions.
• To develop a matrix factorization-based method named MDIPA for this purpose.
• To validate the efficacy of MDIPA in identifying potential miRNA-drug interactions.

Main Methods:

• Developed the microRNA-drug interaction prediction approach (MDIPA), a matrix factorization-based method.
• Utilized experimentally validated miRNA-drug interactions, drug similarity, and miRNA similarity.
• Constructed a path-based miRNA similarity matrix and a drug similarity matrix using structural information.

Main Results:

• MDIPA demonstrated superior performance compared to four state-of-the-art methods in independent dataset and cross-validation tests.
• The method effectively predicts unknown interactions among microRNAs and drugs.
• Molecular docking in a breast cancer case study confirmed the efficacy of MDIPA.

Conclusions:

• MDIPA is an effective computational tool for predicting potential microRNA-drug interactions.
• The approach offers a more efficient alternative to experimental identification methods.
• The findings have implications for drug discovery and therapeutic development targeting miRNAs.