Fig. 5: The relationship of THRB with disease progression.

a, Heatmap of estimated transcriptional network (regulon) activity across the SteatoSITE dataset. b, Patterns of activity and cluster membership of the three regulons (THRB, AEBP2, BNC2). c, Relationships of THRB, AEBP1, BNC2 regulatory networks with significant numbers of component genes from the TRS as gene target members in disease progression. Red edges, positive regulation of gene target; blue edges, negative regulation of gene target; blue nodes, net negative gene regulation; red nodes, net positive gene regulation; gray nodes, net neutral gene regulation; green nodes, TRS gene member. d, Ranked THRB regulon differential activity plot confirms the negative relationship between THRB regulon activity and disease stage (left and center), and the Kaplan–Meier plot indicates that lower THRB regulon activity is predictive of hepatic decompensation (right); log-rank test P value. e, In histologically identical high-risk fibrosis stages (F3 and F4), low THRB activity identifies patients at high risk of hepatic decompensation event; log-rank test P value < 0.0001. f, Example individual two-tailed gene set enrichment plots of target genes in the THRB regulon (left, two-tailed gene set enrichment testing with Benjamini and Hochberg (false discovery rate) adjusted P values < 0.001) from two patients with identical F4 stage (cirrhosis) on PSR-stained sections (right), where the patient with low THRB differential regulon activity (top, blue) progressed to a hepatic decompensation event 224 days after biopsy in contrast to a patient with high THRB regulon activity who did not experience a decompensation event during the 4,500 days until censoring (bottom, red). Scale bars, 2 mm. dES, differential enrichment score.