Features of Genetic Sigma Factors Learning Model Simulating Neural Network

Abstract

Sigma factors play a crucial role in the gene regulation process, which bind with RNA polymerase to unwind the genesequence by identifying the recognition of transcription starting motif pattern, a combination of nucleobases (Adenine(A),Cytosine(C), Guanine(G), Thymine(T)). The advancements in DNA analysis are useful for the geneticist to learn differentpatterns from different perspectives for the identification of mutations in genetic structures, new organisms ranging fromunicellular to multicellular, and useful for creating new gene patterns to get relief from hereditary diseases. To meet all thesechallenging needs, the proposed research work aimed to predict the DNA motif patterns of various sigma factors. Thus themain objective is to create a novel method named "Features of Genetic Sigma Factors Learning Model simulating NeuralNetwork" to predict the prefix motif patterns of major sigma factors such as sigma 70, sigma 32, sigma 24, sigma 19, and sigma38(σ70, σ32, σ24, σ19, and σ38). Each of the sigma factors possesses significant functionality, like vegetative growth for thedevelopment of nutrients. In the proposed model, the novel idea is a generation of a dictionary of DNA motifs that mimics then-gram of natural language processing. The proposed model is trained to feed the DNA motif dictionary, which consists ofpositive and negative motif patterns. The model is tested by an array of K-mer motif patterns taken from the whole E.Colibacterial genome, downloaded from the NCBI website (Escherichia coli str. K-12 substr. MG1655, complete genomeACCESSION: NC_000913). The experimental results of the proposed model yielded 100% accuracy. The model's outcome isa set of patterns that are highly helpful to experts in the biological fields to identify new gene patterns.