Identification of Genes Using Improved Approach Based on GenomeScan and Blast Algorithms
Authors: Bhat, H.F., Wani, M.A. and Rasool, M.
Journal: 12th Indiacom 5th International Conference on Computing for Sustainable Global Development Indiacom 2018
Pages: 3482-3487
Abstract:To understand the biological functions of the DNA or protein sequences, genome annotation proves to be very helpful. The prediction of the genes is one of the most necessary phases of genome annotation. Hence, to predict or locate the gene patters in DNA sequences, several programs have been developed. However, identification of genes is still a difficult problem. There is a rapid increase in data sequence. Hence, the capability of predicting genes in sequences is a challenging task. This paper first reviews the problem of gene prediction and its challenges. This is then followed by the brief analysis of various methods of gene prediction. We have introduced an enhanced way for detecting genes which allows merging of the outputs of two gene prediction programs i.e. GenomeScan and BLAST. This improved method and other existing methods are evaluated on HMR195 dataset and Burset-Guigo dataset. The performance of these programs is measured on the basis of sensitivity and specificity at the nucleotide base level. It is shown that the combined method of gene prediction produces better results than other programs implemented individually. At last, the drawbacks of these methods are detailed which concludes that enhancements are required and need to consider the future directions.
Source: Scopus