Portal de Periódicos da CAPES

PremipreD: Precursor miRNA Prediction by Support Vector Machine Approach

2018; Volume: 11; Issue: 1 Linguagem: Inglês

10.3923/tb.2018.17.24

2077-2254

Sasti Gopal Das, Hirak Jyoti Chak, Abhijit Datta,

Cancer-related molecular mechanisms research

Background and Objective: Precursor microRNA expressions vary depending on their cellular environment and a large amount of genome segments can be folded in similar pseudo precursorʼs microRNA hairpins like structure.Therefore, detection of true precursor microRNA in a genome is challenging task.The computational prediction of precursor MicroRNAs first distinguishes a large amount of similar folded hairpins like structure in genome sequence as a pseudo or true precursor miRNAs.However, researchers need to be improving methods for identification of precursor MicroRNA in a genomic sequence.Materials and Methods: In this computational method, supervised machine learning approach was used as a classifier for classifying the true precursor miRNAs using sequence and secondary structure information.Results: The support vector machine (SVM) classifier achieved accuracy (Q) of 96.28% for predicting true pre-miRNAs.Here, a new precursor miRNA identification tool-PremipreD was developed which performs better in comparison to existing tools, in terms of overall performance and specificity.Conclusion: The PremipreD algorithm reduces the number of false positive prediction rate by using effective Support vector machine methods.