Network Stability Collapse in Lung Adenocarcinoma Reveals a Critical Transition at r* = 0.27

Cancer progression can be understood as a critical transition in the underlying molecular interaction network, where coordinated regulation collapses into fragmented, dysregulated states. In this dataset and accompanying analysis, we present a quantitative network-based stability framework applied to The Cancer Genome Atlas (TCGA) Lung Adenocarcinoma (LUAD) cohort (N = 540), revealing a sharp global connectivity collapse at a refined correlation threshold r* = 0.27. We constructed gene co-expression networks across a continuous range of Pearson correlation thresholds and quantified global network stability using the normalized size of the largest connected component (SG), interpreted as a global stability indicator. As the correlation threshold increases, weak and intermediate interactions are progressively removed, simulating loss of regulatory coordination. We observe a pronounced phase-transition-like collapse in network connectivity, where SG rapidly declines from near-unity to near-zero over a narrow threshold interval. This behavior identifies a refined critical threshold r* = 0.27, representing the first crossing point of accelerated stability loss. This critical point marks the transition between a highly integrated, system-level coordinated regime and a fragmented, disconnected regime, analogous to percolation transitions in complex systems. The result suggests that the LUAD transcriptomic network operates near a critical boundary, where relatively small perturbations in correlation structure lead to disproportionate losses in global connectivity. The dataset includes: The computed global stability indicator curve SG(r) The refined critical threshold identification Visualization of network topology before and after collapse The analysis code and generated figures This framework provides: A quantitative biomarker of global transcriptomic stability A reproducible pipeline for critical threshold detection A systems-level indicator of cancer-associated network destabilization The findings support the hypothesis that cancer can be characterized as a network-level phase transition, and that the refined critical threshold may serve as a candidate biomarker for systemic loss of regulatory coordination. This resource enables further investigation into critical phenomena in biological networks and provides a reference implementation for stability-based network analysis in cancer systems biology