A probabilistic framework to improve microrna target prediction by incorporating proteomics data

Area of Science:

• Molecular Biology
• Bioinformatics
• Genomics

Background:

• Experimental identification of microRNA (miRNA) targets is difficult and low-throughput.
• Existing computational methods primarily rely on sequence information, leading to low overlap and high false-positive rates.
• MicroRNAs primarily function by repressing mRNA translation.

Purpose of the Study:

• To develop a more accurate computational model for predicting miRNA targets.
• To integrate sequence complementarity, miRNA expression, and protein abundance for improved prediction.
• To investigate the dominant mechanism of miRNA-mediated post-transcriptional regulation.

Main Methods:

• Developed a probabilistic model integrating sequence complementarity, miRNA expression levels, and protein abundance to predict miRNA targets.
• Used a second probabilistic model to analyze mRNA expression of confident targets.
• Investigated miRNA-mediated post-transcriptional regulation mechanisms.

Main Results:

• The integrated probabilistic model identified a set of confident miRNA targets.
• Analysis revealed that translational repression, not mRNA degradation, is the predominant miRNA regulatory mechanism.
• This finding explains the previously observed discrepancies between mRNA and protein expression levels.

Conclusions:

• A novel probabilistic approach enhances the accuracy of miRNA target prediction.
• MicroRNA regulation predominantly occurs via translational repression.
• Understanding miRNA mechanisms clarifies the relationship between mRNA and protein abundance.