ABSTRACT
The present study introduces a novel algorithmic approach for identifying lipase-producing and non-lipase-producing yeast using yeast genomic data. A yeast expression (transcriptome) dataset, retrieved from the Gene Expression Omnibus database, is utilized for training an artificial neural network (ANN) incorporated with a genetic algorithm to identify the genes responsible for lipase production and classify lipase-producing yeast cells. The processed gene expression data is used to develop a robust ANN model to classify lipase-producing yeast (LPY) and non-lipase-producing yeast (NLPY). The over-expressed gene acts as a critical factor in classification; thus, it forms a component of the robust model. In the unsupervised learning process to classify LPY and NLPY, the developed model achieved an accuracy of 87.77% with a mean square error of 170. The learning process reduced the total cost (or) loss function value to 0.12. Concerning the futuristic scope of the research, this finding would help classify and identify various biologicals with their expression data.
