Gene Prediction of Prokaryotic Genome Using Neural Network: Array

The need for gene prediction has been increasing throughout the years since the beginning of Human Genome Project. With the emergence of next-generation sequencing technologies, the sequencing of Deoxyribonucleic Acid (DNA) and Ribonucleic Acid (RNA) sequence can be done within lesser time and money. However, the annotation process is unable to keep up with the enormously growing sequence. Thus, a lot of DNA sequences still remains as characters as they have yet to be deciphered by scientist by analyzing and determining the location and the function of the gene within the sequence. However, finding a gene from a vast amount of character manually is nearly impossible and this is where gene prediction becomes helpful. Gene prediction allows the scientists to narrow down the scope from the DNA sequence to be further researched by predicting potential coding region in the sequence using computational methods. Researchers from Malaysia Genome Institute have recently sequenced the genome of Enterobacter sp. but the sequence has not been annotated yet. Therefore, artificial neural network is proposed to determine the position of the potential gene to save the researchers’ effort and time to find gene from Enterobacter sp. An artificial neural network (ANN) will be implemented in order to classify whether a certain region is a potential coding region. The DNA sequence from the Enterobacter sp. genome provided by Malaysia Genome Institute will be used as the subject data of the artificial neural network. It is concluded that the length of the genome does not have any significant correlation to the gene prediction performance while the minimum length of the ORF extracted as candidate for neural network have a significant to the gene prediction.