An unbiased survey of distal element-gene regulatory interactions with direct-capture targeted Perturb-seq

We developed a new framework for highly powered, unbiased CRISPR screens of distal regulatory elements, including an optimized experimental method (Direct-Capture Targeted Perturb-seq (DC-TAP-seq)), a random design strategy, and a comprehensive analytical pipeline that accounts for statistical power. Using pilot datasets from K562 and WTC11 cells, we benchmarked DC-TAP-seq against conventional scRNA-seq and TAP-seq for capture efficiency, guide detection, and knockdown assessment. Applying this optimized framework, we systematically perturbed ~1,400 randomly selected candidate regulatory elements across the two cell types and revealed key features of the human regulatory landscape. We found that most regulatory interactions have small effect sizes (<10%), occur over short genomic distances (<100 kb), and often involve CTCF-bound elements lacking classical enhancer marks. Housekeeping genes were regulated by distal elements with ~2-fold weaker effect sizes compared to other genes. Our results highlight limitations in existing CRISPR datasets and predictive models, and demonstrate the utility of DC-TAP-seq for building more comprehensive maps of distal regulation in the human genome. Overall design: We generated pilot datasets from K562 and WTC11 cells to develop and optimize DC-TAP-seq, a targeted single-cell RNA-seq method for capturing both gene expression and guide RNAs. To benchmark assay performance, we also generated conventional scRNA-seq and TAP-seq datasets from K562. We use the pilot datasets for comparisons of capture efficiency, guide detection, and knockdown assessment.