Fig 2: Cohesin distribution is minimally changed in Stag1-/- and Stag2-/- mESCs. a Western blot analysis of cohesin subunit levels in wild-type, Stag1-/-, and Stag2-/- nuclear lysates. b Genome browser tracks showing STAG1 signal in wild-type and Stag2-/- cells, and STAG2 signal in wild-type and Stag1-/- cells. c MA plots showing differential enrichment of STAG1 signal between wild-type and Stag2-/- cells. Differential enrichment of STAG2 signal between wild-type and Stag1-/- cells is also shown. d Average signal plots of STAG1 signal at CTCF sites, enhancers, and promoters in wild-type and Stag2-/- cells. STAG2 signal at CTCF sites, enhancers, and promoters in wild-type and Stag1-/- cells is also shown. e Genome browser tracks showing RAD21 signal in wild-type, Stag1-/-, and Stag2-/- cells at Stag1-/- and Stag2-/- differential peaks. f MA plots showing differential enrichment of RAD21 in Stag1-/- and Stag2-/- cells. g Average signal plots of RAD21 signal at CTCF sites, enhancers, and promoters in wild-type, Stag1-/-, and Stag2-/- cells

Fig 3: Dual loss of STAG1 and STAG2 reveals redundant functions. a Genome browser tracks for RAD21 ChIP-seq signal in the four conditions: wild-type siGLO, wild-type siStag1, Stag2-/- siGLO, and Stag2-/- siStag1 cells. b Average signal plots and heatmaps of RAD21 signal at a union peak list from the four conditions. c Venn diagram of DEGs from Stag1-/- siGLO, Stag2-/- siGLO, and Stag2-/- siStag1 cells. d Clustered heatmap of log2 fold changes for a combined list of DEGs in wild-type siStag1, Stag2-/- siGLO, and Stag2-/- siStag1 cells all relative to wild-type siGLO. e Violin plots depicting log2-fold changes of all DEGs, and those within Super-enhancer Domains and Polycomb Domains for wild-type siStag1, Stag2-/- siGLO, and Stag2-/- siStag1 cells all relative to wild-type siGLO. Dotted lines indicate the mean. Asterisks indicate significant differences between groups (****p < 0.0001). f Bar graphs with log2 fold change of expression of cell identity genes including those that represent pluripotency (Pou5f1, Sox2, Nanog), ectoderm (Pax6 and Nestin), and endoderm lineages (Gata6 and Sox17) in wild-type siStag1, Stag2-/- siGLO, and Stag2-/- siStag1 cells, all relative to wild-type siGLO. Asterisks indicate significant differences from wild-type siGLO cells (padj < 0.01). g Bar graph with log2 fold change of expression of genes inside and outside of Super-enhancer Domains in wild-type siStag1, Stag2-/- siGLO, or Stag2-/- siStag1 cells, all relative to wild-type siGLO. Asterisks indicate significant differences from wild-type siGLO cells (padj < 0.01). h Western blot analysis following co-immunoprecipitation of CTCF and cohesin in the four conditions. i Proliferation assay with relative cell number represented as a fraction of original plating density for the four conditions. Asterisks indicate significant differences between groups (****p < 0.0001, **p < 0.01, *p < 0.05). Significance at 72 h is between wild-type siStag1 and Stag2-/- siStag1 conditions

Fig 4: Overlapping distribution of cohesin-STAG1 and cohesin-STAG2 on the genome. a Genome browser tracks showing ChIP-seq signal for STAG1, STAG2, RAD21, and CTCF (Z-score normalized). b Average signal plots of STAG1 and STAG2 signal at STAG1 peaks and STAG2 peaks (Z-score normalized). c Frequency of peaks overlapping known functional elements in the genome: CTCF sites, enhancers, promoters, or other (none of the above). d Clustered heatmaps displaying STAG1, STAG2, RAD21, and CTCF signal (Z-score normalized) at a union list of STAG1 and STAG2 peaks. e Western blot analysis following co-immunoprecipitation of SMC1, STAG1, and STAG2 from nuclear lysates

Fig 5: PDS5 proteins localize to the same genomic sites as STAG proteins. A Clustered heatmaps of PDS5A, PDS5B, STAG1, STAG2, and RAD21 ChIP-seq signal at a union list of PDS5A, PDS5B, STAG1, and STAG2 peaks, ordered by RAD21 signal (k-means = 2) (Z-score normalized). B Western blot analysis following co-immunoprecipitation of IgG (negative control), PDS5A, PDS5B, and RAD21 in WT nuclear lysates, under low and high stringency conditions. Control blots for this experiment are in Fig. 1E. C Clustered heatmap of log2 fold change in expression for a combined list of DEGs in Pds5a-/- siGLO, Pds5a-/- siStag1, Pds5a-/- siStag2, and Pds5a-/- siPds5b mESCs relative to WT siGLO mESCs. D Violin plot of log2 fold change in expression for all DEGs, those within Super-enhancer Domains, and those within Polycomb Domains for Pds5a-/- siGLO, Pds5a-/- siStag1, Pds5a-/- siStag2, and Pds5a-/- siPds5b mESCs relative to WT siGLO mESCs. Significance was determined using a Kruskal–Wallis test followed by Dunn’s multiple comparisons test. Asterisks indicate significant differences between groups (* p < 0.05, **** p < 0.0001). E Bar graphs of log2 fold change in expression of pluripotency genes (Pou5f1, Sox2, Nanog), ectodermal lineage genes (Pax6 and Nestin), and endodermal lineage genes (Gata6 and Sox17) in Pds5a-/- siGLO, Pds5a-/- siStag1, Pds5a-/- siStag2, and Pds5a-/- siPds5b mESCs. Asterisks indicate significant differences from WT siGLO mESCs determined using DESeq2 (padj < 0.01)