Alzheimer's disease diagnosis based on gene selection and cross-tissue validation.

Background: Early and accurate diagnosis is critical for intervening in the progression of Alzheimer's disease (AD). However, traditional clinical diagnostic techniques, such as invasive cerebrospinal fluid punctures and expensive neuroimaging, struggle to meet the demands of large-scale early screening. Methods: This study proposes a novel transcriptomic-based diagnostic framework named DeepFS, which features an automated gene selection module to filter out redundant noise and identify high-value core gene targets for diagnosis. Results: DeepFS achieved high-precision identification of AD, yielding an ac-curacy of 89.45% and an AUC value of 0.9647 on brain tissue transcriptomic data. To further explore the methodological generalizability of the framework across different tissue types, we extended DeepFS to a peripheral blood transcriptomic dataset, where it consistently achieved reliable diagnostic classification performance. Furthermore, regarding the brain tissue transcriptomic data, the high-weight genes driven by the gene selection module were integrated with AD brain single-nucleus RNA sequencing data for quantitative analysis at single-cell resolution, further validating the biological efficacy of the module. Conclusion: DeepFS not only exhibits competitive performance in AD diagnostic classification, but can also serve as a valuable prior reference for future AD biomarker research.