Codon based back propagation neural network approach to classify hypertension gene sequences

American Heart Association published a report on 17
^th
December, 2014 where hypertension is narrated as one of the foremost causes of death annually related to heart diseases. So it is a vital issue to predict hypertension from gene sequence to diagnosis this disease initially. According to this prediction, we can prevent and do treatment accurately. In the field of bioinformatics, various machine learning methods like Support Vector Machine (SVM) or Artificial Neural Network (ANN) are used to analyze and classify gene sequences. The Back Propagation Neural Network is very effective to classify the hypertension gene sequence and identify the disease with 90% accuracy rate for small number of sample (80). This paper proposed a method to classify hypertension gene sequence using BPNN after extracting the features from the specified gene sequence. Frequencies of codons (nucleotide triplets) have been used to specify gene sequences. The performance of BPNN technique in training and testing phase has been analyzed for different number of samples. The accuracy of this technique increased according to the number of samples proportionally. Increase of the number of samples reduces the error rate of classification for BPNN classifier.