Emerging Drug Combinations for Targeting Tongue Neoplasms Associated Proteins/Genes: Employing Graph Neural Networks within the RAIN Protocol

Abstract

AbstractBackgroundTongue Neoplasms is a common form of malignancy, with squamous cell carcinoma of the tongue being the most frequently diagnosed type due to regular mechanical stimulation. Its prevalence remains on the rise among neoplastic cancer cases. Finding effective combinations of drugs to target the genetic and protein elements contributing to the development of Managing Tongue Neoplasms poses a difficulty owing to the intricate and varied nature of the ailment.MethodIn this research, we introduce a novel approach using Deep Modularity Networks (DMoN) to identify potential synergistic drug combinations for the condition, following the RAIN protocol. This procedure comprises three primary phases: First, employing Graph Neural Network (GNN) to propose drug combinations for treating the ailment by extracting embedding vectors of drugs and proteins from an extensive knowledge graph containing various biomedical data types, such as drug-protein interactions, gene expression, and drug-target interactions. Second, utilizing natural language processing to gather pertinent articles from clinical trials involving the previously recommended drugs. Finally, conducting network meta-analysis to evaluate the comparative efficacy of these drug combinations.ResultWe utilized our approach on a dataset containing drugs and genes as nodes, connected by edges indicating their associated p-values. Our DMoN model identified Cisplatin, Bleomycin, and Fluorouracil as the optimal drug combination for targeting the human genes/proteins associated with this cancer. Subsequent scrutiny of clinical trials and literature confirmed the validity of our findings. Additionally, network meta-analysis substantiated the efficacy of these medications concerning the pertinent genes.ConclusionThrough the utilization of DMoN as part of the RAIN protocol, our method introduces a fresh and effective way to suggest notable drug combinations for addressing proteins/genes linked to Tongue Neoplasms. This approach holds promise in assisting healthcare practitioners and researchers in pinpointing the best treatments for patients, as well as uncovering the fundamental mechanisms of the disease.HighlightsA new method using Deep Modularity Networks and the RAIN protocol can find the best drug combinations for treating Tongue Neoplasms, a common and deadly form of cancer.The method uses a Graph Neural Network to suggest drug pairings from a large knowledge graph of biomedical data, then searches for clinical trials and performs network meta-analysis to compare their effectiveness.The method discovered that Cisplatin, Bleomycin, and Fluorouracil are suitable drugs for targeting the genes/proteins involved in this cancer, and confirmed this finding with literature review and statistical analysis.The method offers a novel and powerful way to assist doctors and researchers in finding the optimal treatments for patients with Tongue Neoplasms, and to understand the underlying causes of the disease.Abstract Figure