Fig. 3: TF perturbations in mosaic organoids reveal critical regulators of neurodevelopmental fate decisions.
From: Inferring and perturbing cell fate regulomes in human brain organoids
a, Schematic of the single-cell TF perturbation experiment using the CRISPR droplet sequencing (CROP-seq) method. b, The minimum–maximum-scaled average expression (log[transcript counts per 10,000 + 1]) of targeted genes in NPCs, IPs and neurons of the primary and organoid cortex. c, The proportion of cells with each perturbation for each experiment. d, UMAP embedding with cells coloured by detected gRNA (left) and branch assignment (right). e, Regional enrichment of gRNAs. The sidebar shows the number of gRNAs that were consistent and the circles represent consistent effects between experiments and statistically significant (FDR < 0.01) effects on composition. The arrows indicate the predominant observed effect. f, UMAP embedding coloured by consistent gRNAs for selected genes that had a strong effect on fate regulation. g, The Spearman correlation of HES1-target (top, n = 18 genes) and GLI3-target (bottom, n = 42 genes) genes to transition probabilities into the dorsal branch. The GRN was subsetted to retain connections that are accessible at the branchpoint (>5% detection rate). The centre line represents the median, the box limits show the 25–75% interquartile range and the whiskers indicate 1.5× the interquartile range. h, Schematic of the GLI3 loss-of-function experiment using an inducible CRISPR–Cas9 nickase system. i, UMAP embedding of scRNA-seq data from 6-week-old WT and GLI3-KO brain organoids showing the trajectories from NPCs to neurons coloured by different clusters assigned to regional branches. The inset is coloured by genetic condition. j, Stacked bar plots showing the distribution of cluster (colour) assignment per organoid for each condition. k, Differential expression (DE) in ventral telencephalic neurons for GLI3-KO data and CROP-seq data containing a GLI3 gRNA. The x and y axes indicate the coefficients of the linear model. Colours indicate significance (FDR < 10−4) in CROP-seq, the KO cell line or both.
