Machine learning method for the Prediction of glaucoma associated Genes and their differential Expression analysis through Bioinformatics tools

Abstract

The human gene expressions are the relative factors that determine newlinethe ability of the protein to cause disease. Predicting the proteins ability to newlinecause glaucoma by analysing its amino acid composition and arrangements newlinemight play a crucial role in fighting many eye related diseases. Though the newlineavailability of many specialized database is committed to provide the protein newlinesequence information, the limitation in practical exercises exists in newlinedetermining the proteins that are associated with glaucoma. Glaucoma is a newlineneurodegenerative disease and second leading cause of blindness worldwide. newlineThe disease is characterized by an elevated intraocular pressure. Carbonic newlineanhydrase plays a major role by forming aqueous humor and its inhibition can newlinereduce intraocular pressure by partially suppressing the secretion of aqueous newlinehumor. In this scenario, the development of computational methods such as newlinemachine learning approaches serves as time consuming alternatives. Thus, in newlinethis paper the classifier methods were evaluated to understand its ability in newlineterms of sensitivity, specificity and accuracy in predicting the protein as newlineglaucoma associated or not. This study revealed that SMO classifier (Support newlinevector machine) excels in predicting the proteins association with Glaucoma newlinedisease based on its Amino Acid Composition (AAC) at five-fold newlinecross-validation. Predicting the protein sequence as glaucoma plays a key role newlinein the development of novel drug to combat with eye cancer. In this study, newlinevarious classifier methods available in data mining techniques were evaluated newlineon the non-redundant dataset of 4086 (2035 glaucoma and 2051 newlinenon-glaucoma) protein sequences retrieved from swissprot database while newlineusing amino acid composition as input features newline