Operon Prediction Using Broad Learning System

An operon is a fundamental unit of transcription and contains specific functional genes for the construction and regulation of network at the entire genome level. Operon prediction becomes important for us to study to discover the mystery of life. By predicting the operon we can use at the unknown cell to predict the operon and genes to find out it function and regulatory system. The methods normal use for operon prediction are intergenic distance, participation in the same metabolic pathway, the cluster of orthologous groups and the phylognetic profile are use as the input for the Broad Learning System(BLS). The reason why we use BLS to do this prediction problem is because the accuracy, sensitivity and specificity of this problem done by other comparative method is not enough high. We can get the raw data source of Escherichia coli genome easier but we need to extract out the features as the input by difference way. The dataset we use have total 4242 rows so we split 50% to training sets and 50% testing sets in which means that we have both 2121 training and testing as the input of BLS. This method has achieve our expectation which reach a high accuracy, sensitivity, and specificity which are 100%, 100%, and 100% respectively for the genome in a complex form of data.