Integrative gene co-expression network analysis reveals protein-coding and LncRNA genes associated with Alzheimer’s disease pathology

Abstract

Alzheimer’s disease (AD) is a complex neurodegenerative disorder marked by widespread molecular changes, many of which remain poorly understood. While AD pathology progresses through specific brain regions, it is unclear whether these regions are affected similarly. Long non-coding RNAs (lncRNAs), emerging as key cellular regulators, remain largely uncharacterized in AD. Understanding how lncRNAs interact with protein-coding genes across brain regions could shed light on AD mechanisms and progression. To investigate this, we performed consensus weighted gene co-expression network analysis on 396 postmortem brain RNA-seq samples using a meta-analytic approach. Our analysis revealed substantial network rewiring in AD, particularly in the temporal cortex compared to the frontal cortex. The temporal cortex exhibited adaptive changes in gene interactions, while the frontal cortex showed a breakdown of healthy correlations—possibly reflecting regional differences in disease progression. We identified 46 protein-coding genes and 27 lncRNAs as key components in the AD network of the temporal cortex. Using known functions of protein-coding genes as reference points, we inferred potential functions for over 100 lncRNAs across both regions. These findings highlight novel lncRNA candidates potentially involved in AD and provide insights into their roles in both healthy and diseased brain states.