Targeted Perturb-seq enables genome-scale genetic screens in single cells

Abstract

The transcriptome contains rich information on molecular, cellular, and organismal phenotypes. However, experimental and statistical limitations constrain sensitivity and throughput of genetic screening with single-cell transcriptomics readout. To overcome these limitations, we introduce targeted Perturb-seq (TAP-seq), a sensitive, inexpensive, and platform-independent method focusing single-cell RNA-seq coverage on genes of interest, thereby increasing the sensitivity and scale of genetic screens by orders of magnitude. TAP-seq permits routine analysis of 1,000s of CRISPR-mediated perturbations within a single experiment, detects weak effects and lowly expressed genes, and decreases sequencing requirements up to 50-fold. We apply TAP-seq to generate perturbation-based enhancer-target gene maps for 1,778 enhancers within 2.5% of the human genome. Thereby, we show that enhancer-target association is jointly determined by 3D contact frequency and epigenetic states, allowing accurate prediction of enhancer targets throughout the genome. In addition, we demonstrate that TAP-seq can identify cell subtypes with only 100 sequencing reads per cell.

Transcriptional enhancers: from prediction to functional assessment on a genome-wide scale

“…Applications of single cell RNA-seq to pooled CRISPR screens, including those causing multienhancer perturbation, may better account for complex and potentially redundant enhancer interactions to offer better reproducibility in phenotype detection (Sanjana et al 2016;Schraivogel et al 2020). Understanding the versatile roles of enhancers in transcriptional regulation may offer the exciting possibility of therapeutic enhancer editing in situations where enhancer mutations alter gene expression and lead to an altered phenotype (Kvon et al 2020).…”

“…Although the CRISPR-Cas9 nuclease system has been a boon for large knockout screens of coding sequences, these screens have fallen short for screening entire collections of predicted regulatory elements until recently. To overcome issues related to the throughput and sensitivity of single cell RNA-seq following genetic perturbation, the targeted Perturb-seq (TAPseq) protocol was developed (Schraivogel et al 2020). By greatly decreasing the sequencing requirements for single-cell transcriptomics following CRISPR enhancer perturbation, 1778 predicted enhancers were targeted and 81 enhancer-target gene pairs were functionally defined in a human erythroleukemia cell line (Schraivogel et al 2020).…”

Gene network modeling via TopNet reveals robust epistatic interactions between functionally diverse tumor critical mediator genes

“…In contrast, network modeling based on perturbation experiments allows one to directly assess cause-effect relationships before creating epistatic maps. This has led to increased interest in perturbation experiments as a means of modeling biological systems, such as the use of Perturb-seq 36,37,38,39 to measure the effect of perturbations at single-cell resolution. cDNA clones were obtained from the IMAGE consortium collection, distributed by Open Biosystems, or PCR-cloned from murine cDNA using sequence-specific primers except for murine Jagged2 (Jag2) and Notch3-intracellular domain (Notch3-ICD) (gifts of Dr. L. Milner).…”

“…Several computational and statistical challenges will need to be addressed in this context, such as the increased feature space, decreased sensitivity, and unidirectional perturbations. In the near term, targeted Perturb-seq39 , in which some of these challenges are lessened, is ideally suited for initialmethods development and testing. Our general modeling approach has the potential to extract rich biological interactions from these data and open new avenues of biological inquiry.…”

Interpretable Machine Learning for Perturbation Biology

“…Larger datasets can be obtained by measuring diverse types of molecular data, for example, transcriptomic, epigenomic, and metabolomic changes (Brown et al , 2014;Zaal and Berkers, 2018) . A major opportunity for larger datasets may arise from recent cell barcoding techniques that significantly increase perturbation throughput relative to arrayed experiments (Adamson et al , 2016;Dixit et al , 2016) by measuring levels of transcripts by sequencing, levels of proteins detected by antibodies labeled by oligonucleotides or isotopes at the single cell level, or levels of both by multiplexing (Frei et al , 2016;Wroblewska et al , 2018;Mimitou et al , 2019;Schraivogel et al , 2020) . A key advantage of single cell perturbation approaches for data-driven inference of dynamic networks is scale, as barcoded perturbation experiments can be pooled and individual cells identified based on sequence tags.…”