GKLOMLI: a link prediction model for inferring miRNA-lncRNA interactions by using Gaussian kernel-based method on...

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Limited understanding of microRNA-lncRNA interactions hinders the elucidation of regulatory mechanisms in human diseases.
• Experimental validation of these interactions, such as CLIP-seq, is costly and time-consuming.
• Computational tools are increasingly vital for predicting potential miRNA-lncRNA interactions for further experimental validation.

Purpose of the Study:

• To develop a novel computational model for predicting miRNA-lncRNA interactions.
• To provide a reliable and efficient tool for identifying potential miRNA-lncRNA interactions.
• To facilitate the study of gene expression modulation in human diseases.

Main Methods:

• Proposed a novel link prediction model named GKLOMLI (Gaussian kernel-based method and linear optimization algorithm for inferring miRNA-lncRNA interactions).
• Utilized a Gaussian kernel-based method to generate miRNA and lncRNA similarity matrices from an observed interaction network.
• Employed a linear optimization algorithm trained on an integrated matrix for predicting interactions.

Main Results:

• The GKLOMLI model demonstrated high performance in predicting miRNA-lncRNA interactions.
• Cross-validation experiments, including k-fold CV and leave-one-out CV, yielded high Area Under the Curves (AUCs).
• Achieved AUCs of 0.8623 ± 0.0027 (2-fold CV), 0.9053 ± 0.0017 (5-fold CV), 0.9151 ± 0.0013 (10-fold CV), and 0.9236 (LOO-CV).

Conclusions:

• The proposed GKLOMLI method is a high-performance tool for predicting miRNA-lncRNA interactions.
• GKLOMLI is expected to aid in uncovering underlying interactions and deciphering mechanisms of complex diseases.
• This computational approach offers a valuable resource for future biological experiments and disease research.