MicroRNAs (miRNAs) have twenty two nucleotides and possess small non-coding single-stranded RNA molecules which are used for gene expression at the post-transcriptional level1. Their influence on gene regulation is much prominent; therefore the value of miRNA research is being amplified in molecular biology. They are linked with numerous gene regulation phenomena, including those of neurodegenerative disorders, diabetes, cancer, cell development as well as cell death2. Moreover, discovery of new miRNAs is still a huge chanllenge; therefore, a computational approach is needed to examine the miRNA expression and for miRNA detection3.
It is reported that pre-miRNA hairpin secondary structure as well sequential information has the potential in the computational identification of precursor miRNAs4. Machine learning procedures like Support Vector Machine (SVM) are useful in this regard. Most of these approaches employ known pre-miRNA as a positive data set and pseudo hairpins as a negative dataset to train their models.
The vast number of sequences in the genome can fold into miRNA precursor like hairpin secondary structures5. The computational prediction of pre-miRNAs first distinguishes a genome sequence as a pseudo or true precursor miRNA6. But, scientists are utilizing incomplete techniques to identify precursor miRNAs in genomic sequences.
Therefore, Sasti Gopal Das, Hirak Jyoti Chakraborty and Abhijit Datta7 designed an experiment for this purpose, in which a machine learning approach based was presented on the support vector machine to identify the precursor miRNA genes. They employed a reliable computational approach (PremipreD) for better identification of precursor miRNAs. In this computational technique, supervised machine learning approach was utilized as a classifier to classify the true precursor miRNAs by means of sequence as well as secondary structure information.
At the end of this experiment, scientists found that the PremipreD classifier can predict new precursor miRNA from animal, plant as well as viruses. Conclusively, the comparison of prediction capabilities between PremipreD and other existing procedures, regarding specificity and overall accuracy specify that the overall performance of the PremipreD algorithm considerably is better as compared to all other tools.
Your email address will not be published. Required fields are marked *
17 November, 2019