Incorporation Local Protein Structure Information With Clustering Based Algorithm To Predict Structural Classes

The structural class’s prediction using local protein structure information approach according to three different classes namely for all-?, all-? and mixed class will increase the performance measurement. Moreover, the selection of SSAM and different length of local protein structure will impact in achieving higher success rate in which giving a more reliable and quality prediction. The classification rule used in this research has contributed to a remarkable success rate against CATH. The usage of clustering based technique proves that advantage of K-means in solving the cases of non-identical structural class using the optimal number of clusters is promising. With the capability of predicting structural classes by local protein structure will hence provide guidance and assistance in predicting protein global structure and finally identify protein function. In the field of biology, physicians can to identify the sickle protein or to design the specified 72 medicine or novel drugs in curing diseases and illnesses. This will be enable an early detection of patients will fatal disease such as cancer by researchers. The sufficient sequence-structure knowledge also cater various opportunities in computational method that can inspire more innovative solution in the near future.