Background Single-cell CRISPR screening (sc-CRISPR) is a functional genomics method combining pooled CRISPR (sgRNA) genetic screening with single-cell expression analysis. Our study aims to compare the performance of three different sc-CRISPR screening methodologies using the 10X Genomics scCRISPR platform: 10X-3’ scCRISPR, CROP-Seq (using modified 10X-3’ scCRISPR protocol), and 10X-5’ scCRISPR. With 10X-3’ scCRISPR, the CS1 capture sequence inserted in the sgRNA tracr is used to detect the expressed sgRNAs. With the CROP-Seq protocol the sequence of expressed sgRNAs is detected within the poly-dT cDNA pool; with the 10X-5’ scCRISPR method, sgRNAs are detected without needing to be either modified with the insertion of the CS1 sequence or embedded within the mRNA transcript.

Methods To test the performance of the three different sc-CRISPR approaches, we ran three parallel sc-CRISPR screens with an ~200 sgRNA library focused on a small subset of genes associated with TNFα response, cloned in a 10X-CS1 lentivector, CROP-Seq lentivectors, or unmodified sgRNA lentivector, respectively. HEK293-Cas9 cells were transduced with the three lentiviral libraries at MOI ~0.05, puromycin-selected to enrich for transduced cells, treated for 72h with TNFα, and analyzed at the single-cell level for the activation of the TNFα response pathways.

Results Our results show that while the CROP-Seq and 10X-CS1 approaches are equally effective for single-cell sgRNA detection and gene expression profiling, the CROP-seq approach was superior insofar as the sgRNA’s ability to induce the expected phenotype. These findings were consistent with anecdotal observations that the CS1-modified sgRNAs are less active than unmodified sgRNAs. The 10X-5’ scCRISPR protocol, however, proved to be superior to both CROP-Seq and 10X-CS1 in sgRNA detection, and comparable to CROP-seq in gene expression profiling and sgRNA activity. We also explored the possibility of modifying the 10X-5’ scCRISPR protocol, to use a targeted sc-RNA-seq approach instead of the standard transcriptome-wide RNA-seq. We reasoned that this modification could lead to lower NGS costs and better transcriptional profiling results. Unfortunately, while the targeted sc-CRISPR RNA-seq approach proved to be technically successful, it did not seem to provide superior transcriptional profiling results, and the savings on the NGS costs were counterbalanced by the cost and effort associated with the design of the targeted RNA-seq primer detection panel.

Conclusions These insights provide a valuable framework for selecting a suitable sc-CRISPR screening method and highlight the 10X-5’ scCRISPR protocol as a promising technique for robust sgRNA detection and gene expression analysis.