Immunotherapy is becoming a promising strategy for treating diverse cancers. However, it benefits only a selected group of gastric cancer (GC) patients since they have highly heterogeneous immunosuppressive microenvironments. Thus, a more sophisticated immunological subclassification and characterization of GC patients is of great practical significance for mRNA vaccine therapy. This study aimed to find a new immunological subclassification for GC and further identify specific tumor antigens for mRNA vaccine development. First, deep autoencoder (AE)-based clustering was utilized to construct the immunological profile and to uncover four distinct immune subtypes of GC, labeled as Subtypes 1, 2, 3, and 4. Then, in silico prediction using machine learning methods was performed for accurate discrimination of new classifications with an average accuracy of 97.6%. Our results suggested significant clinicopathology, molecular, and immune differences across the four subtypes. Notably, Subtype 4 was characterized by poor prognosis, reduced tumor purity, and enhanced immune cell infiltration and activity; thus, tumor-specific antigens associated with Subtype 4 were identified, and a customized mRNA vaccine was developed using immunoinformatic tools. Finally, the influence of the tumor microenvironment (TME) on treatment efficacy was assessed, emphasizing that specific patients may benefit more from this therapeutic approach. Overall, our findings could help to provide new insights into improving the prognosis and immunotherapy of GC patients.

Chemotherapy is a fundamental therapy in cancer treatment, yet its effectiveness is often undermined by drug resistance. Understanding the molecular mechanisms underlying drug response remains a major challenge due to tumor heterogeneity, complex cellular interactions, and limited access to clinical samples, which also hinder the performance and interpretability of existing predictive models. Meanwhile, single-cell RNA sequencing (scRNA-seq) has emerged as a powerful tool for uncovering resistance mechanisms, but the systematic collection and utilization of scRNA-seq drug response data remain limited. In this study, we collected scRNA-seq drug response datasets from publicly available web sources and proposed a transfer learning–based framework to align bulk and single cell sequencing data. A shared encoder was designed to project both bulk and single-cell sequencing data into a unified latent space for drug response prediction, while a sparse decoder guided by prior biological knowledge enhanced interpretability by mapping latent features to predefined pathways. The proposed model achieved superior performance across five curated scRNA-seq datasets and yielded biologically meaningful insights through integrated gradient analysis. This work demonstrates the potential of deep learning to advance drug response prediction and underscores the value of scRNA-seq data in supporting related research.