Decoding neuronal gene expression: integrative insights from omics and AI

Neuronal functional diversity and pathological vulnerability are governed by multi-layered regulatory programs. While high-throughput omics and neuroimaging provide high-resolution snapshots of these programs, bridging the gap between molecular dynamics and macro-scale brain architecture remains a significant informatics challenge. This review synthesizes the evolution of computational frameworks in neuro-omics-transitioning from descriptive co-expression modules to causal graph neural networks and cross-scale foundation models. We evaluate these methodologies within the context of Alzheimer's disease, schizophrenia, and epilepsy, identifying critical bottlenecks in data harmonization, spatial alignment, and causal interpretability.