Identification of Nuclear Receptors and Its Family Using Wavelet Variance Features

Nuclear receptors are important transcriptional regulators involved in widely diverse physiological functions such as control of embryonic development, cell differentiation, and homeostasis. In addition, these molecules are extremely important in medical research since a large number of them are implicated in diseases such as cancer, diabetes, or hormone resistance syndromes. Identification of nuclear receptors and further classifying it into its family from its primary sequence is an important problem in computational biology. In this paper, we propose a novel feature vector based on wavelet variance of seven important physicochemical properties of amino acids. Furthermore, the dimension of the proposed feature vector is equal to 35. This helps in building faster and memory efficient classifier than dipeptide composition based approach. Experimental results show that the proposed approach identifies nuclear receptors from non-nuclear receptors with an accuracy 99.91% using a 10-fold cross-validation technique. The proposed technique is further able to identify six families nuclear receptors with an overall accuracy of 96.19% using a 5-fold cross-validation technique. Tests were also performed on unseen datasets for nuclear receptors super- family and its family.