An Algorithm for Plant MicroRNA Target Prediction

Abstract

MicroRNAs (miRNAs) are RNAs of~24-nucleotide in length which by binding to the 3' untranslated region (3'-UTR) of the target mRNA, bearing complementary target sequences or degrading mRNA by cleaving at single site, cause translational suppression and thereby down.regulate gene expression. The study of the relationship between miRNAs and their target mRNAs is now an attractive area in bioinformatics. Predicting miRNAs, which target mRNAs using experimental methods is a challenging task as it is time consuming and costly. In the present study, for predicting the target sequences in mRNAs, a computational method namely “miRNA target plot†was developed using an efficient R program involving data input, target prediction and plotting. The mature miRNA sequence and specific mRNA sequence information were entered and the sequence information was read using seqinR package. Input data sequences were further processed in two steps. In the first step, the user input miRNA sequence as fasta format in 5'-3' direction was reversed in 3'- 5’direction using function rev.comp ( ) in R package sequinR and the complement of this sequence was used for finding the optimal match based on sequence similarity. In the second step, the program cuts the mRNA sequence from the first position onwards till the end of mRNA sequence and equals the length akin to the length of miRNA. All the possible alignments between each miRNA-mRNA pair was determined using dynamic programming and the scores were calculated. The positive scores were filtered out and the optimal target sequence was found. The miRNA and mRNA segment alignment scores and mRNA segment positions were also plotted in a scatter plot. The computational method is equally effective in predicting target mRNAs compared with other existing tools like TAPIR, psRNAT which has been verified with sequences from StarBase database. The miRNA target plot can be used for precisely predicting target mRNAs for miRNAs.