BACKGROUND: Prediction of the oral resistome protein sequence is crucial to have multi-drug efflux pumps for both drug discovery and resistance management. Sophisticated machine learning and deep learning techniques called graph attention networks (GATs) and hyperattention processes are used to predict protein sequences and their interactions in molecular biology. Our goal is to forecast the oral multi-drug resistance efflux protein sequence's graph attention network. METHODS: The FASTA sequences (UniProt ids: P0ACR9, P0ACH5, P60752, P0AE82, P0A7V8, P24177, P76399) were retrieved from MprA, MarA, CpxA, and MdtC. These sequences were located and their quality assessed. The Deepbio program was used to analyze hyperattention and graph attention networks using FASTA sequences. RESULTS: Sensitivity = TP / (TP + FN) = 0.51 represents the percentage of positive cases identified by hyper attention and graph attention network, while the percentage of negative cases the model properly recognizes is 0.83 TNR. also known as particularity. Technique for computation: Hyper attention and the graph attention network have specificities of TN / (TN + FP) = 0.83 and 0.91, respectively. CONCLUSIONS: Oral multi-drug resistome sequences can be predicted using the GAT model, which can also be applied to future resistant sequences.
YADALAM et al. (Fri,) studied this question.