Evolution Scale Modelling for Prediction of Regulator G Protein Signalling (RGS) Sequences in Oral Cancer

Oral cancer is a major global health concern, causing significant morbidity and mortality. Despite advancements in diagnostics and treatments, the 5-year survival rate remains low, highlighting the need for novel biomarkers and therapeutic targets. Understanding the molecular mechanisms of oral carcinogenesis is crucial for improving patient outcomes. The G protein-coupled receptor (GPCR) signaling cascade is pivotal in regulating various cellular functions, including cell proliferation, differentiation, and apoptosis. RGS proteins, which regulate the G protein signaling network, are crucial in cancer progression, including oral cancer and their dysregulation is linked to various diseases. This study uses evolutionary scale modeling to predict RGS protein sequences for treating oral cancer.The data was extracted from the UniProt protein sequence database, normalized, and divided into 80% training and 20% testing. This meticulous process ensured high-quality data for the ESM model, which is crucial for understanding protein data complexities. The model’s pre-training on extensive protein datasets improves its performance on specific tasks.The Evolution Scale Modelling (ESM) protein model underwent training and validation, showing a significant decrease in initial loss and a steady improvement in accuracy. The model’s overfitting index is low, indicating good generalization. The learning rate behavior is controlled, and the model achieves high precision and recall in classification tasks. Its versatility in various scenarios is evident.The study showcases the effectiveness of Evolutionary Scale Modeling (ESM) in predicting RGS protein sequence roles in oral cancer, achieving high training and validation accuracy, precision-recall performance, and stability indicators.