Supplementary Figure 3: Identification of consensus sets of Bcl11b target genes.

(a), Experimental scheme is shown. BM-derived precursors from Cas9-Bcl2-tg mice were cultured on OP9-DL1 for 7 days, then they were infected with sgRNA. Seven days after infection, CD25⁺ sgRNA transduced cells were purified and subjected to RNA-seq analysis. (b), Overlaps between gene sets differentially expressed when Bcl11b is disrupted by Lck-Cre or Vav1-iCre in vivo, compared with those differentially expressed when Bcl11b is deleted acutely by Cas9 and sgRNA in bone marrow cell precursors in vitro. See Supplementary Table 1 and 2 for rpkm values and gene lists. Conditions for deletion by Cas9 are described in detail below (Fig. 5) and in Methods, and rpkm values are given in Supplementary Table 3. (c), Overlaps between gene sets differentially expressed by Bcl11b deletion in the present study with gene sets previously reported to be differentially expressed by Bcl11b deletion in fetal liver-derived precursors differentiating in vitro⁷. DEGs affected by Lck-Cre deletion in vivo, Vav1-iCre deletion in vivo, and Cas9-mediated deletion in vitro in the present study (“consensus set”) are compared with those previously reported using retroviral Cre transduction to delete Bcl11b in fetal liver precursors (“Gold standard” minimal sets of responding genes in Longabaugh et al.⁷). DEG numbers in the earlier study are lower because fetal liver-derived precursors differentiate faster and the conditions used gave less consistent differentiation extents, reducing statistical significance. The consensus set includes a much higher fraction of the previously reported targets than those genes affected by Cas9-mediated deletion in vitro but not affected in vivo (a control for cell culture effects). Note that agreement among Bcl11b-repressed DEGs in all four sample types is higher than agreement among Bcl11b-dependent DEGs.