Efficient combinatorial targeting of RNA transcripts in single cells with Cas13 RNA Perturb-seq

In singlet cells, RNA and bcgRNA labels showed 99.2% concordance.

“Moreover, we observed extremely high concordance between RNA and bcgRNA labels in singlet cells (99.2%) (Fig. 1d,e).”

Transcriptomic pseudobulk profiles for all 26 targeting gRNA groups showed decreased mRNA expression for each targeted transcript, including when three genes were perturbed simultaneously.

“When examining transcriptomic pseudobulk profiles for all 26 targeting gRNA groups, we observed decreased messenger RNA expression for each targeted transcript, even when perturbing transcripts of three genes simultaneously (15 examples in Extended Data Fig. 3f,g).”

Indirect guide capture with an optimized configuration gave the strongest crRNA transcript detection among the tested methods, for both catalytically active and inactive Cas13 proteins.

“While all methods successfully induced robust knockdown (Fig. 1b), we found that indirect guide capture with an optimized configuration resulted in the strongest crRNA transcript detection ability (configuration X, Fig. 1c), for both catalytically active and inactive Cas13 proteins.”

Knockdown strength was similar for multi-gRNA crRNA arrays compared with single-gRNA perturbations.

“Moreover, the strength of knockdown was similar for multi-gRNA crRNA arrays relative to single gRNA perturbations (Fig. 2b and Extended Data Fig. 3c–e).”

Both synergistic and dampening effects were observed among dual perturbations.

“We also observed both synergistic and dampening effects.”

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.

Cell cycle distributions did not differ between perturbed cells and NT control cells.

“Last, we classified cell cycles stage for each cell, and found no difference in cell cycle distributions (indicative of their proliferation state) compared to NT control cells (Extended Data Fig. 4d–f).”

Alternative Cas9-based approaches struggled to efficiently identify and detect combinatorial perturbations compared with CaRPool-seq.

“Our benchmarking analysis found that, in contrast to CaRPool-seq, alternative Cas9-based approaches struggled to efficiently identify and detect combinatorial perturbations (Fig. 3b).”

Differentially regulated gene modules were typically associated with genetic programs of myeloid cell differentiation and function.

“These differentially regulated gene modules were typically associated with genetic programs associated with the differentiation and function of myeloid cells (Extended Data Fig. 8c).”

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.

Fitting regression models to pairwise perturbations revealed a diversity of genetic interactions that clustered into four groups.

“Fitting these models to each of our pairwise perturbations revealed a diversity of genetic interactions, which we broadly clustered into four groups (Fig. 4g).”

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.

Observed log2 fold changes for all perturbations were reproducible across technical replicate perturbations using independent gRNAs (R = 0.82).

“Observed log2 fold changes for all perturbations were also reproducible (R = 0.82, Extended Data Fig. 7f) across technical replicate perturbations when comparing effects measured for independent gRNAs.”

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.

The CaRPool-seq experiments did not observe direct or indirect evidence for RfxCas13d promiscuous collateral activity.

“Lastly, while we did not observe direct or indirect evidence for RfxCas13d’s promiscuous collateral activity in our CaRPool-seq experiments, other RNA-targeting CRISPR effectors37,40,41 may represent an alternative for future experiments.”

• cites [37] Özcan, A. et al. Programmable RNA targeting with the single-protein CRISPR effector Cas7-11. Nature 597, 720–725 (2021). sourceNot checked yet

  Not checked in the free run.

• cites [40] Tong, H. et al. High-fidelity Cas13 variants for targeted RNA degradation with minimal collateral effects. Nat. Biotechnol. https://doi.org/10.1038/s41587-022-01419-7 (2022). sourceNot checked yet

  Not checked in the free run.

• cites [41] Wei, J. et al. Deep learning and CRISPR-Cas13d ortholog discovery for optimized RNA targeting. Preprint at bioRxiv https://doi.org/10.1101/2021.09.14.460134 (2022). sourceNot checked yet

  Not checked in the free run.

Eight genes with potential off-target binding sites for the three perturbations showed no significant expression change in the CaRPool-seq data.

“We identified eight genes whose sequence contains potential off-target binding sites for any of our three perturbations, but found that the expression of these genes was not significantly changed in our CaRPool-seq data (Extended Data Fig. 4a).”

Differential gene expression signatures for single-gene perturbations from ECCITE-seq (5′ scRNA-seq, Cas9) were readily reproduced in CaRPool-seq (3′ scRNA-seq, Cas13d) data.

“We found that the differential gene expression signatures for single-gene perturbations obtained using ECCITE-seq (5′ scRNA-seq, Cas9 perturbation)23 can be readily reproduced in our CaRPool-seq (3′ scRNA-seq, Cas13d perturbation) data (Extended Data Fig. 8a,b).”

• cites [23] Wang, E. et al. Surface antigen-guided CRISPR screens identify regulators of myeloid leukemia differentiation. Cell Stem Cell 28, 718–731 (2021). sourceNot checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.

Transcriptional responses varied widely when KDM1A knockdown was paired with different chromatin regulators.

“We generally observed that transcriptional responses varied widely when pairing KDM1A knockdown with different chromatin regulators.”

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.

In KRAB-dCas9-MeCP2 THP1 cells, dual perturbation of KDM1A and HDAC3 synergistically upregulated CD11b and CD14 surface markers after 7 days.

“To further validate the synergistic relationship between KDM1A and HDAC3 we infected KRAB–dCas9–MeCP2 expressing THP1 cells with multiple independent sgRNAs and measured CD11b and CD14 cell surface protein expression 7 days postinfection using flow cytometry. Again, we observed synergistic upregulation of both surface markers (Fig. 4d) after dual perturbation.”

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.
• AssumptionRests on an assumption: CD14 and CD11b are immunophenotypic markers of monocyte differentiation.

When EP300 and KDM1A were combinatorially perturbed, the EP300 signature appeared more strongly than the KDM1A signature.

“For example, we found that the EP300-signature appeared more strongly than the KDM1A-signature when combinatorial perturbing both genes (Extended Data Fig. 8g).”

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.

The increased sensitivity systematically improved the signal-to-noise ratio for distinguishing true bcgRNA UMI counts from spurious secondary bcgRNA UMI counts.

“The increased sensitivity also systematically improved the signal-to-noise ratio to distinguish true bcgRNA UMI counts from spurious secondary bcgRNA UMI counts (Extended Data Fig. 9d).”

Many cluster 1 pairs represented perturbations of two proteins in the same complex (e.g., MED14/MED24, SUPT16H/SUPT6H), which showed similar gene expression responses for single and double perturbations.

“As a positive control, many of the pairs in this cluster represented perturbations of two proteins in the same complex (that is, MED14/MED24, SUPT16H/SUPT6H), which show similar gene expression responses for singly and doubly perturbed cells (Extended Data Fig. 8d–f).”

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.

Cluster 1 also included pairs of proteins in separate complexes (MED24/SMARCD1) that share similar perturbation signatures.

“This cluster also represented pairs of proteins residing in separate complexes (MED24/SMARCD1 of mediator and SWI/SNF complexes), which share similar perturbation signatures (Extended Data Fig. 8d).”

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.

Mitochondrial gene expression was not upregulated in cells receiving one or more targeting gRNAs compared with NT controls in the CaRPool-seq data.

“Therefore, we compared mitochondrial gene expression levels in cells that have received one or more targeting gRNAs compared to NT control cells (Extended Data Fig. 4c), and did not observe upregulation in our CaRPool-seq data.”

When KDM1A was paired with RING1 perturbation, the KDM1A response signature dominated.

“In contrast, the KDM1A response signature dominated when paired with perturbation of the polycomb repressive complex member RING1 (Extended Data Fig. 8h).”

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.

Individual perturbations of selected target genes led to detectable CD11b expression changes after 3 days.

“Through individual perturbations with a flow-cytometry readout, we found that target gene perturbations led to detectable CD11b expression changes after 3 days (Extended Data Fig. 6i).”

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.
• AssumptionRests on an assumption: CD14 and CD11b are immunophenotypic markers of monocyte differentiation.

The transcriptional response to dual KDM1A/HDAC3 perturbation placed cells at later segments of the integrated myeloid differentiation trajectory, showing enhanced differentiation compared with either single perturbation.

“This transcriptional response led dually perturbed cells to be distributed at later segments of our integrated myeloid differentiation trajectory, exhibiting enhanced differentiation compared to either single perturbation (Fig. 4i,j).”

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.

The evopreQ1 pseudoknot element led to sixfold higher bcgRNA detection UMI counts compared with the initial design.

“In particular, the evopreQ1 pseudoknot element led to a sixfold higher bcgRNA detection UMI counts compared to our initial design (Extended Data Fig. 9c).”

In the KRAB-dCas9-MeCP2 experiment, of 1,570 cells that received vectors targeting three genes, only 779 (49.6%) were associated with the correct three sgRNAs after sequencing.

“For example, in the KRAB–dCas9–MeCP2 experiment, we recovered 1,570 cells that received vectors targeting three genes. Among these cells, only 779 (49.6%) were associated with the correct three sgRNA after sequencing.”

In cluster 4, genetic interactions were identified where one gene's effect appeared to dominate over the other.

“In cluster 4, we identified genetic interactions where one gene’s effect appeared to dominate over the other.”

• AssumptionRests on an assumption: The THP1 MLL-AF9 NRASG12D cell line is a valid model system for AML differentiation.

Bulk RNA-seq after CD55 knockdown found no evidence for elevated Cas13d off-targets.

“We also performed CD55 target knockdown followed by more sensitive bulk RNA-seq and found no evidence for elevated levels of Cas13d off-targets (Extended Data Fig. 4b and Supplementary Table 3).”

When considering cells with three successfully detected perturbations, all methods induced similarly strong depletion of all three surface proteins (Cas13d 74.5%, Cas9 75.5%, KRAB-dCas9 75.2%, KRAB-dCas9-MeCP2 77.3%).

“When considering these cells, all methods successfully induced a similarly strong depletion of all three surface proteins (Cas13d 74.5%, Cas9 75.5%, KRAB–dCas9 75.2%, KRAB–dCas9–MeCP2 77.3%) (Fig. 3c and Extended Data Fig. 5c,d).”

The three targeting approaches (Cas13d, Cas9-nuclease, KRAB-dCas9-MeCP2) produced different numbers of differentially expressed genes in addition to CD55 reduction: 1 for Cas13d, 3 for Cas9, and 30 for KRAB-dCas9-MeCP2.

“The three approaches show a varying number of differentially expressed genes in addition to CD55 reduction (n = 1 Cas13d, n = 3 Cas9, n = 30 KRAB-dCas9-MeCP2).”

Direct capture with pA-handle shows strongly reduced knockdown efficiency compared to regular gRNAs.

“Direct capture with pA-handle shows strongly reduced knockdown efficiency compared to regular gRNAs (standard condition).”

Cell cycle stage assignment did not differ significantly between NT controls and populations targeting varying numbers of genes or using increasing numbers of gRNAs.

“Statistical testing for differences in cell cycle stage assignment for cell populations targeting e) varying numbers of target genes per cell, or using f) increasing numbers of gRNAs per cell did not show significant differences to NT control cell populations expressing zero targeting gRNAs”

Indirect capture constructs show strong target gene knockdown similar to regular gRNAs for all three target genes.

“Indirect capture constructs show strong target gene knockdown similar to regular gRNAs (standard condition) for all three target genes.”

Direct and indirect Cas13 guide RNA capture allow robust target knockdown.

“Extended Data Fig. 1 | Direct and indirect Cas13 guide RNA capture allow for robust target knockdown.”

Across 26 differential gene expression analyses, none of the 13 mitochondrial genes showed significant expression differences from control cells (adjusted p < 0.05).

“Across all differential gene expression analyses (n = 26), none of the 13 mitochondrial genes encoded on the mitochondrial chromosome was expressed significantly different from control cells (adjusted p-value < 0.05; Wilcoxon’s rank sum test).”

Significantly differentially regulated genes other than the target genes had a median expression change of 12.8% (s.d. 9.9%).

“Genes highlighted in black showed a median expression change of 12.8% (s.d. 9.9%).”

All 28 target genes showed consistent gRNA enrichment or depletion.

“Consistent gRNA enrichment/depletion of all 28 target genes.”

CD14 was the most depleted gene, as expected.

“As expected, CD14 was the most depleted gene.”

• cites Wang et al.Not checked yet

  Not checked in the free run.

Differences in target gene knockdown between varying numbers of gRNAs per array are larger than differences between gRNA identities at the same total gRNA count, assuming comparable gRNA efficiencies.

“The Analysis suggests that target gene knockdown differences between the number of gRNAs per array are more pronounced than differences between gRNA identities with the same total gRNA count, given that gRNA efficiencies are comparable as shown in (c).”

MALAT1 and NEAT1 (MENß) sequences required nucleotide exchanges to remove potential terminator sequences and enable full transcription by RNA polymerase III.

“Sequences found in MALAT1 and NEAT1 (MENß) required nucleotide exchanges (shown in red) to remove potential terminator sequences (≥ 4U) and allow the sequences to be fully transcribed by RNA polymerase III.”

Indirect capture constructs achieve strong knockdown of all three target genes, comparable to standard gRNAs.

“Indirect capture constructs show strong target gene knockdown like regular gRNAs (standard condition) for all three target genes.”

Reverse transcription and amplification succeeded for Type-X crRNAs captured indirectly with a smallRNA PCR-handle and polyA-tail.

“Indirect capture of Type-X crRNAs with smallRNA PCR-handle and polyA-tail (arrow) allowed for reverse transcription and amplification.”

GFI1 g2 did not lead to strong effects in the pooled screen and CaRPool-seq experiment.

“We noticed GFI1 g2 did not lead to strong effects in the pooled screen and in the CaRPool-seq experiment.”

• Also stated in: § CaRPool-seq data analysis · plain

  “Comparing differential gene expression results for technical replicates embedded in the CaRPool-seq library, we noticed GFI1 g2 did not lead to upregulation of CD11b ADT or upregulation of the expected gene expression signature.”

Despite potential off-target binding, no transcriptomic perturbation was observed for these genes.

“In Extended Data Fig. 4a, we demonstrate that despite the potential for off-target binding, we do not observe transcriptomic perturbation for these genes.”

• AssumptionRests on an assumption: gRNA matches shorter than 17 bases do not cause target knockdown.

In CaRPool-seq, 74.7% of cells were assigned a single bcgRNA.

“total cells n = 9,355, cells with single bcgRNA n = 6,986, (74.7%)”

The evopreQ1 element increased bcgRNA detection sensitivity by an average of 6-fold (n = 4 per condition).

“The evopreQ1 element yielded on average 6-fold higher bcgRNA detection sensitivity (n = 4 per condition).”

Cas13-mediated knockdown shows a unimodal response, suggesting limited cell-to-cell differences in target gene knockdown.

“Importantly, the Cas13-mediated function shows a unimodal response, suggesting limited cell-to-cell differences in target gene knockdown.”

Knockdown via direct capture using a polyA-tail sequence is limited.

“Target knockdown with direct capture through a polyA-tail sequence is limited.”

All but two gene pairs showed elevated CD11b levels.

“All but two gene pairs show elevated levels of CD11b.”

When two or more perturbations are performed simultaneously, the fraction of cells in which all perturbations are both successfully introduced and successfully detected can decrease dramatically.

“Therefore, when performing two or more simultaneous perturbations, the fraction of cells where all perturbations are both successfully introduced and successfully detected can decrease dramatically.”

The indirect detection strategy is well suited for delivering multiple gRNAs into one cell together with a detectable bcgRNA that encodes their collective identity.

“Notably, our strategy for indirect detection is well-suited for delivering multiple gRNAs into a single cell alongside a detectable bcgRNA that encodes the collective identity of these perturbations.”

Compared with existing Cas9-based methods, CaRPool-seq has a distinctive advantage in efficiently profiling cells that receive combinatorial perturbations.

“We compared CaRPool-seq to existing Cas9-based methods, highlighting its unique strength to efficiently profile combinatorially perturbed cells.”

Undetected or incorrectly assigned single-guide RNAs can affect up to 20% of cells.

“For example, undetected or incorrectly assigned single-guide RNAs can affect up to 20% of cells6; this is compounded when multiple independent sgRNAs are introduced and independently detected in each cell.”

• cites [6] Replogle, J. M. et al. Combinatorial single-cell CRISPR screens by direct guide RNA capture and targeted sequencing. Nat. Biotechnol. 38, 954–961 (2020). sourceNot checked yet

  Not checked in the free run.

Using a unique reverse transcription handle and Illumina PCR priming sequence in the modified crRNA allows perturbations to be detected alongside scRNA-seq and additional modalities such as CITE-seq.

“In addition, using a unique set of reverse transcription handle and Illumina PCR priming sequence in our modified crRNA (Extended Data Fig. 2) ensures that these perturbations can be detected not only alongside scRNA-seq, but also when profiling additional molecular modalities (for example, CITE-seq12 for simultaneous transcriptome and surface protein profiling).”

• cites [12] Stoeckius, M. et al. Simultaneous epitope and transcriptome measurement in single cells. Nat. Methods 14, 865–868 (2017). sourceNot checked yet

  Not checked in the free run.

CaRPool-seq places multiple perturbations on a cleavable CRISPR array linked to a detectable barcode, which allows several genes to be targeted at once.

“CaRPool-seq encodes multiple perturbations on a cleavable CRISPR array that is associated with a detectable barcode sequence, allowing for the simultaneous targeting of multiple genes.”

RfxCas13d crRNAs can be modified by adding a common reverse transcription handle either directly to the gRNA or as a separate bcgRNA in a CRISPR array, enabling reverse transcription and amplification.

“These results demonstrate that RfxCas13d crRNAs can be modified by adding a common reverse transcription handle either directly to the gRNA or as a separate bcgRNA as part of a CRISPR array, allowing for reverse transcription and amplification.”

• AssumptionRests on an assumption: Cas13 crRNAs lack a common priming site for reverse transcription.

CaRPool-seq enables pooled perturbation screens to be efficiently and accurately demultiplexed into a single-cell readout.

“These numbers demonstrate that CaRPool-seq enables pooled perturbation screens that can be efficiently and accurately demultiplexed into a single-cell readout.”

RfxCas13d can process a CRISPR array into multiple mature CRISPR RNAs.

“Notably, RfxCas13d is also capable of processing a CRISPR array into multiple mature CRISPR RNAs (crRNAs)10, presenting an attractive option for combinatorial perturbations at the RNA level.”

• cites [10] Konermann, S. et al. Transcriptome engineering with RNA-targeting type VI-D CRISPR effectors. Cell 173, 665–676 (2018). sourceNot checked yet

  Not checked in the free run.

The strategy is well suited for combinatorial perturbations whose identity is encoded in a single barcode reliably detected alongside multiple modalities including scRNA-seq and CITE-seq.

“We demonstrate that this strategy is well-suited for performing combinatorial perturbations, whose identity is encoded in a single barcode that can be reliably detected alongside multiple molecule modalities including scRNA-seq and CITE-seq.”

Benchmarking shows CaRPool-seq is more efficient and accurate than Cas9-based technologies when assigning multiple perturbations in single cells.

“Through benchmarking, we show that CaRPool-seq is more efficient and accurate when assigning multiple perturbations in single cells when compared to Cas9-based technologies.”

Because CaRPool-seq associates combinatorial perturbations with a single bcgRNA, its detection efficiency does not vary between single and multiple perturbations.

“Since CaRPool-seq associates combinatorial perturbations with a single bcgRNA, the efficiency of detection does not vary between single and multiple perturbations.”

The modified bcgRNA structure further improves CaRPool-seq performance.

“This modified bcgRNA structure therefore further improves the performance of CaRPool-seq, and is recommended for future experiments.”

The selected gRNAs had no secondary target sites with 0-2 mismatches to the cognate site.

“Selected gRNAs had no secondary target sites with 0–2 mismatches to the cognate site42.”

• cites [42] Guo, X. et al. Transcriptome-wide Cas13 guide RNA design for model organisms and viral RNA pathogens. Cell Genomics 1, 100001 (2021). sourceNot checked yet

  Not checked in the free run.

The absence of observed collateral activity likely reflects controlled delivery of Cas13d and gRNAs via single-integration lentiviral systems rather than transient overexpression.

“Our findings likely reflect CaRPool-seq’s controlled delivery of Cas13d and gRNAs using single integration lentiviral systems instead of transient overexpression, which is consistent with the suggestion that lower and more controlled Cas13 expression can mitigate collateral activity16.”

• cites [16] Kelley, C. P., Haerle, M. C. & Wang, E. T. Negative autoregulation mitigates collateral RNase activity of repeat-targeting CRISPR-Cas13d in mammalian cells. Cell Rep. 40, 111226 (2022). sourceNot checked yet

  Not checked in the free run.

Cas13 cleavage may produce RNA molecules that are detectable by scRNA-seq but cannot be translated into functional protein, so measured RNA knockdown may underestimate the phenotypic effect of Cas13 perturbation.

“Further, analogous to how Cas9-nuclease targeting often produces RNAs degraded by nonsense mediated decay13, it is possible that Cas13 cleavage produces RNA molecules that can be detected by scRNA-seq but cannot be translated into functional protein14, suggesting that the level of measured RNA knockdown underestimates the phenotypic effect of Cas13 perturbation.”

• cites [13] Tuladhar, R. et al. CRISPR-Cas9-based mutagenesis frequently provokes on-target mRNA misregulation. Nat. Commun. 10, 1–10 (2019). sourceNot checked yet

  Not checked in the free run.

• cites [14] Burris, B. J. D., Molina Vargas, A. M., Park, B. J. & O’Connell, M. R. Optimization of specific RNA knockdown in mammalian cells with CRISPR-Cas13. Methods 206, 58–68 (2022). sourceNot checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: Target RNAs are continuously produced and degraded before translation.

The CaRPool-seq library had correct gene pair to bcgRNA linkage greater than 94%.

“Complete library representation with minimal bias (90th percentile/10th percentile crRNA read ratio 2.6/4.8), and correct gene pair to bcgRNA linkage (>94%) was verified by Illumina sequencing (MiSeq).”

The optimized strategy delivers multiple gRNAs as part of a single CRISPR array that is subsequently cleaved into individual crRNAs.

“We introduced an optimized strategy to deliver multiple gRNA as part of a single CRISPR array, which is subsequently cleaved into individual crRNAs.”

CaRPool-seq is a flexible method for CRISPR-Cas13 RNA-targeting screens with a single-cell sequencing readout.

“Here, we present CaRPool-seq, a flexible method for performing CRISPR–Cas13 RNA-targeting screens with a single-cell sequencing-based readout.”

Structured RNAs can improve bcgRNA detection in CaRPool-seq.

“Structured RNAs can improve bcgRNA detection in CaRPool-seq experiments.”

Using CRISPR arrays with multiple gRNAs against the same target may further enhance target knockdown.

“CRISPR arrays encoding multiple gRNAs against the same target may be used to further enhance target knockdown.”

The results indicate that indirect gRNA capture can be achieved using polyA-tail capture instead of CS1-based capture.

“These results show that indirect gRNA capture can be facilitated with polyA-tail capture as an alternative to CS1-based capture.”

CaRPool-seq enables specific target RNA knockdown.

“CaRPool-seq enables efficient bcgRNA capture and specific target RNA knockdown.”

The observed expression changes are independent of target RNA and gRNA amounts.

“indicating that observed changes are independent of target RNA and gRNA amounts.”

CaRPool-seq enables efficient bcgRNA capture.

“CaRPool-seq enables efficient bcgRNA capture and specific target RNA knockdown.”