ARGai: AI Fusion Approaches for Classification of S. aureus Strains and Identification of Antibiotic Resistant Genes

Staphylococcus aureus (S. aureus) causes bone and joint infections, pneumonia, and bloodstream infections, the most common and fatal of these illnesses. Rapid identification of resistant strains and the antibiotic resistance genes (ARGs) linked to them can enhance the diagnostic process for various infectious illnesses. We introduce a new clinical support system that uses a hybrid multi-stage significant gene identification technique called MRMR-LASSO to select resistant and significant genes from next-generation sequencing (NGS) datasets. It is necessary to identify the S. aureus resistance strain from among thousands of relevant genes in NGS datasets, which are typically complex and comprise many genes with few samples. The curse of dimensionality makes it hard to pick informative genes from many genes. Optimization techniques are employed to resolve the gene selection issue. To select important genes for our ARGai model, we employed a filter in conjunction with an embedding approach. With its combination of machine learning (ML) and deep learning (DL) methods, ARGai can efficiently and accurately identify the important genes in S. aureus with a 96% success rate. Examining the biological relevance of chosen gene activity guarantees that ARGai's classification performance on susceptible and resistant strains is outstanding. Several cutting-edge AI models used for gene selection and strain classification are compared to our ARGai model. By shortening the screening period, this proposed ARGai may aid in subsequent phases of illness treatment, such as disease diagnosis and medication design.