Introduction: MicroRNAs (miRNAs) are small noncoding RNAs that regulate gene expression by mRNA degradation or preventing translation. They function by binding to 3′ untranslated region (3′UTR) on target mRNAs. Nowadays, miRNAs are considered as one of the main regulators of many biological processes in animals including early development, cell differentiation, proliferation, and apoptosis. According to previous researches, miRNAs are highly expressed in central nervous system and involved in many neural functions and diseases. There is also an increasing interest to examine role of different miRNAs in brain physiology and pathological conditions. For functional miRNA analysis, one critical first step is to identify genes targeted by the miRNA. Identification of miRNAs targets by computational analysis has low cost compared to the experimental molecular methods. In the past decade, some algorithms and databases have been designed and developed to help researchers in predicting miRNA target genes. However, reliable microRNA target prediction is an important and still unsolved computational challenge. All of the developed predicted algorithms employ microarray profiling data or crosslinking and immunoprecipitation (CLIP) sequencing data. Although CLIP data is widely employed in the prediction of microRNA target genes but it has been revealed that miRNA binding to target genes do not result necessarily in functional downregulation of the target genes. One of the other strategies in target prediction is identification of transcripts downregulated by overexpression of miRNAs. Method: In this study, we assessed a new computational algorithm known as MirTarget in identification of miRNA targets. This algorithm is accessible via miRDB site (www.miRDB.org). This algorithm combines CLIP binding dataset and miRNA overexpression data to detect miRNA target genes. All datasets including 96 features of miRNAs is combined to design the target model of MirTarget, which is important for both miRNA binding and target functional downregulation. The final model of MirTarget uses both data of microRNA binding and downregulation of target genes to compute a probability score that reflects the statistical assessment of the prediction accuracy for an individual target site. Result: The predicted target genes by MirTarget model are ranked in an output table in which each target gene takes a score from 0-100 range. Genes with score above 50 are possible targets for an individual miRNA and the score of 100 is the target gene with the highest probability as a target gene of the interested miRNA. Conclusion: Based on the results of the prediction of miRNA transcript target by MirTarget model and comparing it to other previous algorithms, we conclude that MirTarget algorithm is more accurate than the other previous models like TargetScan and miRWalk in predicting the correct and functional targets. Finally, we can propose MirTarget algorithm for miRNA target prediction in neuroscience researches.
