Abstract
Enhancers are the short regularity genome regions, which are bounded with proteins to activate the transcription of a specific gene. Enhancers are further categorized into two classes on the basis of its structure and strength, namely strong enhancers and weak enhancers. Owing to technological improvement, huge numbers of DNA sequences are explored in data banks. So the identification of these unprocessed data via traditional methods is challenging due to intricate and vague structures. In addition, only limited numbers of recognized structures are available. In order to tackle the limitations of traditional methods, it is indispensable to adopt the concept of intelligent and machine learning. In this regard, a two-layer automated model is proposed. In the first layer, it discriminates between enhancer and non-enhancer. In case of enhancer, further, the second layer identifies the types of enhancer. DNA sequences are expressed using pseudo dinucleotide composition, pseudo trinucleotide composition and pseudo tetra nucleotide composition. In order to combine the strength of various feature spaces, a hybrid feature space is formed by amalgamating these three feature spaces. After compiling the results, it is observed that support vector machine achieved encouraging results in conjunction with hybrid feature space, which is 77.86 and 65.83% of accuracies on examined datasets. The obtained results exposed that the proposed model performed prominent compared to the existing approaches so far notifiable in the literature. It is realized that the developed model might be more useful and expedient for basic research and academia.