HINN prioritizes multi-omics biomarkers associated with cognitive decline.

HINN achieves the lowest mean normalized MSE in predicting cognitive decline using multi-omics data, outperforming baseline models across multiple assessments.

• Why it matters: Understanding the biological mechanisms underlying cognitive decline is crucial for early diagnosis and targeted interventions, yet integrating diverse omics data remains complex and underexplored.
• What they did: The study developed HINN, a deep learning framework incorporating biologically informed connections across genomic, epigenetic, and transcriptomic layers, utilizing GWAS, methylation, and gene expression data with pathway mappings.
• The result: HINN identified meaningful cross-omics biomarkers, such as SNP rs116557230, CpG cg24041822, and SOCS4 expression, demonstrating its potential to uncover biologically relevant insights into cognitive decline.