Introduction: Gastric cancer, a leading cause of cancer-related death globally, is driven by a combination of environmental, genetic, and epigenetic factors. Single nucleotide polymorphisms (SNPs) have emerged as critical biomarkers for understanding individual susceptibility to this malignancy. The colossal amounts of genomic data necessitate advanced computational tools to discern patterns and predict SNPs that significantly impact the development of gastric cancer. This review delves into the innovative integration of artificial intelligence (AI) methodologies in forecasting gastric cancer-associated SNPs.
Methods: Literature databases were thoroughly scanned to consolidate research that utilized AI-driven algorithms, including but not limited to machine learning and deep learning, for SNP prediction related to gastric cancer. The selected studies were evaluated based on the AI model used, dataset size, SNP identification process, validation strategies, and prediction accuracy.
Results: AI has revolutionized the identification of gastric cancer-associated SNPs, outperforming traditional statistical methods in terms of speed and accuracy. Several studies have employed machine learning models, including support vector machines, random forests, and neural networks, showing prediction accuracies that often surpass 90%. Deep learning methodologies, though in their infancy in this domain, have showcased potential, especially convolutional neural networks (CNN) and recurrent neural networks (RNN). The integration of feature selection methods with AI models has further improved prediction accuracy by identifying pertinent genomic features and reducing computational complexity. While most studies utilize public genomic databases, the emergence of large-scale multi-omics datasets has empowered models to consider gene-gene and gene-environment interactions, providing a holistic view of SNP-gastric cancer associations.
Discussion: Despite the advancements, challenges remain. Balancing the trade-off between model complexity and interpretability, managing imbalanced datasets, and integrating diverse data types are key areas for improvement. Moreover, the translation of AI-predicted SNPs into clinical settings necessitates rigorous validation and a deeper understanding of the functional implications of these SNPs.
Conclusion: AI, with its capacity to process and analyze vast genomic datasets, has demonstrated remarkable potential in predicting SNPs linked to gastric cancer development. Its continued refinement and integration with multi-omics data are set to offer unprecedented insights into the genetic underpinnings of gastric cancer, thereby advancing personalized medicine and therapeutic strategies.
Keywords: Gastric cancer , Artificial Intelligence, Single Nucleotide Polymorphisms (SNPs)