Aging as loss of communication fidelity in gene regulatory networks

Aging involves progressive loss of cellular function, but the organizing principles of this decline remain unclear. We developed an information-theoretic framework that quantifies age-related changes in the fidelity of transcriptional regulation. Modeling transcription factors (TFs) and their target genes (TGs) as a multi-input, multi-output communication system, we computed mutual information (MI) between TFs and TGs from single-cell RNA-seq data across tissues. MI consistently decreased with age, identifying loss of regulatory fidelity as a quantitative hallmark of aging. Network analyses revealed that this decay arises mainly from input mismatch, meaning the TF activity is operating outside the sensitive range for TG activation, rather than corruption of the network itself. Remarkably, in silico knock-ins of a few TFs restored MI to youthful levels, indicating that reinforcing key regulatory nodes can rejuvenate information flow in aged gene regulatory networks.