Fig. 5: Epiregulon identifies drivers of lineage reprogramming.

a Lentiviral constructs used to introduce TFs into LNCaP cells in the reprogram-seq assay. b UMAP representation of 3903 LNCaP cells transduced with virus encoding GATA6, NKX2-1, FOXA1, FOXA2 and mNeonGreen. The cells were infected in individual wells, hashtagged with HTO and then pooled into a single run. c Distribution of HTO tags in each of the clusters. d Motif enrichment in cluster-specific peaks was performed by ArchR using the CisBP motif database. e Chromatin accessibility at neuroendocrine (NE)—and stem cell-like (SCL)-specific regions defined by Tang et al.³⁹ computed by ChromVAR. f Shown are the distribution of HTO tag assignment for GATA6 and NKX2-1, the level of TF gene expression, chromatin accessibility at GATA6- and NKX2-1 binding sites estimated by ChromVAR and TF activity computed by Epiregulon. g Gene expression (left) and Epiregulon-inferred activity (right) of NKX2-1 and GATA6. h Spearman correlation of regulon weights vs. log-fold changes of putative target genes of GATA6 (left) or NKX2-1 (right) with respect to mNeonGreen-infected cells. Shown is the confidence interval of 95%. P-values are calculated from 2-sided t-test. i Each cell was identified by the HTO tag corresponding to the well receiving virus encoding GATA6 or mNeonGreen. A cell was classified into either expressing GATA6 or not based on its TF activity. AUROC - area under the receiver operating characteristic curve. j Same as i but for NKX2-1. k Number of TFs detected in the GRN computed by the different packages. Source data are provided as a Source Data file.