MDA-CF: Predicting MiRNA-Disease associations based on a cascade forest model by fusing multi-source information

Area of Science:

• Biochemistry and Molecular Biology
• Computational Biology
• Genomics

Background:

• MicroRNAs (miRNAs) are crucial regulators of biological processes.
• miRNAs hold potential as biomarkers and therapeutic targets for various diseases.
• Experimental identification of miRNA-disease associations is costly and time-consuming, necessitating computational approaches.

Purpose of the Study:

• To develop a computational method, MDA-CF, for predicting miRNA-disease associations.
• To leverage multi-source information and a cascade forest model for accurate prediction.
• To provide a reliable tool for uncovering novel disease-associated miRNAs.

Main Methods:

• Integrated multi-source information to represent miRNAs and diseases.
• Utilized an autoencoder for dimension reduction and feature space optimization.
• Employed a cascade forest model for predicting miRNA-disease associations.

Main Results:

• MDA-CF achieved an average AUC of 0.9464 on HMDD v3.2 and 0.9258 on HMDD v2.0.
• Demonstrated superior performance compared to previous computational methods.
• Validated high percentages (100%, 86%, 88%) of top predicted miRNAs for colon, breast, and gastric neoplasms.

Conclusions:

• MDA-CF is a reliable computational method for identifying disease-associated miRNAs.
• The tool offers a promising approach for biomarker discovery and disease association studies.
• The source code is publicly available for further research and application.