Fig. 1: Safeguard repressor screen.

a, Schematic representation of safeguard repressor prediction based on TF expression and DNA-binding motif analysis. b, Scores of top safeguard repressor candidates across 18 cell types, including lifelong expression, repressor/activator activity, tumor suppressor roles and TFs promoting indicated cell fate (asterisks). c, TF expression and motif presence analysis highlight top six hepatocyte safeguard repressor candidates. d, Prox1 expression (left) and motif counts in signature genes of indicated cell types (right). e, log-rank test between Kaplan–Meier curves from patients with HCC in TCGA, segregated by high versus low expression of indicated candidates. f, Bulk RNA-seq expression of hepatocyte candidates during mouse liver development17. g, Validation of top three liver candidates by overexpression using 4-in-1 iHep reprogramming. h, TJP1 immunofluorescence of cells in g (n = 3). i, Quantification of TJP1+ cells and albumin secretion of cells in g (n = 3). Scale bar = 100 µm (h). Bar and line graphs show mean (n = 3), and error bars = s.d. (f,i). Two-tailed Dunnett’s test (i). P values are displayed. TCGA, The Cancer Genome Atlas; iHep, induced hepatocyte; RPKM, reads per kilobase million; WB, western blot; IF, immunofluorescence.