Predicting miRNA-disease association through combining miRNA function and network topological similarities based on...

Area of Science:

• Bioinformatics
• Genomics
• Computational Biology

Background:

• Predicting microRNA-disease associations is crucial but existing methods often overlook complex network characteristics.
• Current approaches rely heavily on similarity networks, limiting comprehensive analysis.

Purpose of the Study:

• To develop a novel method, ComSim-MINE, for predicting microRNA-disease associations by integrating function and network topology similarities.
• To address the limitations of existing methods by incorporating complex network characteristics.

Main Methods:

• ComSim-MINE calculates combined similarity using the harmonic mean of miRNA function similarities and network topology similarities.
• The method identifies functional modules within the miRNA functional interaction network.

Main Results:

• ComSim-MINE demonstrates competitive performance against state-of-the-art algorithms like ClusterONE, MCODE, NEMO, and SPICi.
• The method achieved satisfactory results in F-measure, sensitivity, and accuracy on the miRNA functional interaction network.
• Case studies revealed new potential associations, offering clinical insights for diseases such as COVID-19.

Conclusions:

• ComSim-MINE provides an effective approach for predicting microRNA-disease associations by leveraging combined similarities and network topology.
• The findings offer valuable clues for understanding and potentially treating epidemic diseases, including COVID-19.