Single-Cell CRISPR Screening

Single-cell CRISPR screening has provided scientists a level of resolution that arrayed CRISPR screening cannot. Arrayed screening simply cannot interrogate the number of cells that single-cell screens can. When studying the transcriptome on a single-cell level, the addition of genetic perturbation can give scientists a comprehensive understanding of how individual genes affect the entire transcriptome.

The first iterations of single-cell CRISPR screening technologies included methods like Perturb-seq, CROP-seq, and CRISP-seq. These platforms all use modified lentiviral vectors to introduce gRNAs into the cells. Once the cells have been perturbed, it’s a matter of separating these cells using a combination of barcodes, microfluidics, or other techniques, generating sequencing libraries, and analyzing the data. Since the introduction of these methods, there have been many single-cell CRISPR screening methods that have built upon these foundational technologies.

Sequencing the gRNA

One of the main challenges when it comes to single-cell CRISPR screening is linking changes observed in the transcriptome to the perturbation that caused these changes. “It's really important to connect the guide barcode or the gene that you're targeting with the transcriptomic readout, and you need that on a single-cell level,” says Melanie Janczyk, Director of Market Development at Scale Biosciences. However, gRNAs can be hard to capture because they’re lowly expressed. Many of the tools out there for single-cell CRISPR screening are meant to optimize gRNA detection.

Scale Biosciences enhances the sequencing efficiency of gRNAs by modifying the traditional CROP-seq method. To capture the transcriptome and specific guides, “we do both poly A reverse transcription and guide-specific reverse transcription,” explains Janczyk. After this, the method uses a guide-specific PCR to further enhance gRNA detection. “Those two steps are really meant to boost the capture of your guide so that you’re getting the most out of your experiment,” she adds.

Parse Biosciences takes another approach by using combinatorial barcoding. In this method, they take cells that have been already perturbed by a CRISPR library, split them up into a 96-well plate, and add a barcode. Then they pool the cells together and split that into a second plate to receive the next round of barcodes. This process is then repeated two more times.

“As cells go from one pool to another, they follow a unique path in this process,” says Charlie Roco, CTO and Co-founder of Parse Biosciences. “The unique path corresponds to the DNA barcodes at the end of all the molecules inside the cell.” Then a subset of the cells is enriched for guide sequences separately so that two sequencing libraries are generated: one for gRNAs and one for the transcriptome. Because the same barcodes exist per cell between the libraries, they can be paired computationally.

Maximizing throughput

No matter the exact method for single-cell CRISPR screening, it’s important to maximize throughput in the wet-lab portion of the experiment so that you capture as many cells as possible. First, you need good transfection. “The number of cells that you can profile in these types of experiments ends up mattering a lot,” says Roco. “You want to really maximize the number of guide sequences and the number of cells that get a guide sequence.” Janczyk says many of the popular cell lines will reach efficiencies of up to 95% of the cells receiving a guide. For others, such as T cells, it can be lower. She advises doing a small-scale experiment first to see what the guide capture efficiency is before beginning a large-scale screen.

Computational challenges

With more data comes the need for more computational power and tools that can handle large amounts of data. “Where people are running into some bottlenecks as these experiments are getting really big is that the standard bioinformatics tools out there aren’t meant to look at millions of cells at a time,” says Janczyk.

Another computational challenge is to match up the guide RNAs to the perturbations. With Parse Biosciences' method, which generates a guide RNA library and a transcriptome library, matching these might not be as straightforward if there is cross-talk between the cells (i.e., what appears to be the same barcode in different cells). “The most important thing is that you're accurately assigning the guide sequence to the correct cell,” says Roco.

“Once you have clean data, that’s where the fun starts,” notes Roco. “You can start mapping different perturbations to what genes are changing.” Many tools such as mixscape, sceptre, and MUSIC are available online to do this.

Accelerating the drug discovery pipeline

Single-cell CRISPR screening has greatly enabled the drug discovery pipeline. “We're seeing an explosion in the field right now of people trying to do these very large screens,” Roco explains. “They are finding new discoveries and new drug candidates that are making their way to the clinic by doing these early large screens. And they're cutting time dramatically by being able to do this.”

Janczyk is excited to see screens go bigger. She says that with the tools out there now, it’s possible to do “exploratory and discovery science in a way that you weren't able to before because you were limited by technology and cost.” Roco agrees, “the power of single-cell CRISPR screens is high, and it becomes even more important when you scale up and do more cells.”