Using Emerging Subsequence in Classifying Protein Structural Class

Knowledge about protein's structure can help in understanding its function and has many applications in computer-aided drug design and protein engineering. In this paper we introduce a new methodology for predicting protein structural class using Emerging Subsequences (ES). In a sequence database, an emerging subsequence of data class is a subsequence which occurs more frequently in that class rather than other classes. They can capture significant contrast between data classes. Our idea is to discover all the ES from protein sequence database and use as representatives for this data. Our experimental results using CATH database shows good result when evaluating the accuracy of the proposed method.