Research Article - Journal of RNA and Genomics (2022) Volume 18, Issue 4

Convolution neural network for the prediction of RNA using heterogeneous network.

T N6-methyladenosine (m6A), a kind of post-transcriptional alteration, is essential for the stability and control of gene regulations. As a result, identifying m6A is critical for comprehending the functional mechanisms of biological systems. To make the tedious process easier, many machines learning algorithms based on convenient handicraft feature extractions techniques had been presented. Nevertheless, due to poor extracting features, such strategies enhance computing overhead and as a result, reduce the reliability of m6A detection. That research provides a rapid and accurate statistical method for m6A location detection. This suggested approach relies on the CNN, where recovers the much more important aspects from RNA sequences encode by appending as well as nucleotides chemical composition. This proposed approach is tested to state-of-the-art prediction algorithms on different species benchmarks datasets. Here on a benchmark dataset of Homo sapiens (H.sapien), Mus musculus (M.musculs), Saccharomyces cerevisiae (S.cerevisiae), as well as Arabidopsis thaliana (A.thaliana), the proposed system provides good precision of 93.6 percent, 93.8%, 85.0% and 92.5%, correspondingly.

Author(s): P Sivakumar*

Abstract PDF