Since the advent of next-generation sequencing, it has never been quicker, easier, and cheaper to sequence a genome and unlock the information encoded within. But knowing the genotype is only one part of the puzzle—the myriad of genetic information now at our disposal far outstrips our knowledge and understanding of gene function and the roles played in biological processes and disease states.

Functional genomics aims to link genotype with phenotype, but on a genomic scale. Genome-wide pooled library screens utilizing CRISPR-Cas9 gene-editing technology are now increasingly popular, allowing scientists to perform unbiased, systematic genome-wide loss of function studies within the technical capabilities of most laboratories. The simplicity and precision of CRISPR-Cas9 has meant that it has overtaken RNAi as the method of choice for administering genomic screens, and provides an exciting opportunity to elucidate cellular process, understand disease states, and identify targets for drug discovery. But a CRISPR screen involves the selection and optimization of a multitude of steps to ensure that the months of effort and expense result in good data.

Which model system?

The first issue to address when embarking on a CRISPR pooled library workflow involves two conflicting issues—first, the relevance to the biological question to be answered, and second, the practical reality of the chosen system. For example, primary cells, organoids, and animal models would be the preferred option, but they may not be amenable to transduction, and the tens of thousands of sgRNA generally involved in a genome-wide screen would make them impossible to work with at scale. Conversely cell lines would be practically advantageous but perhaps not as biologically relevant, making it harder to validate hits that arise.

In instances where the preferred model system cannot be used with a genome-wide library, a cell line could be used for a primary screen to identify hits, before moving into the more biologically relevant model for follow up. Alternatively, a smaller library could be considered for small-scale screening in the more appropriate model, for example targeting specific protein families such as kinases.

Technical suitability—maintaining library representation

A CRISPR-Cas9 pooled screen involves delivery of a genome-wide sgRNA library by lentiviral vector—the stable integration of the sgRNA cassette provides each cell with a barcode allowing identification of those sgRNA that have been enriched or depleted following application of a selection pressure by PCR amplification and massively parallel sequencing. The viral titer is purposely kept low to ensure only a single integration event per cell—however, a typical workflow is a multi-step process involving transduction with the lentiviral library, selection of infected cells, proliferation, seeding, and passaging followed by isolation of genomic DNA and PCR amplification.

Whichever model system has been selected, optimization of each of these steps must be performed—a time-consuming process but crucial for the success of the screen. An important consideration in this optimization stage is maintaining library representation, preventing loss of sgRNA within the library, as the accurate quantification of sgRNA is the cornerstone of identifying hits. In his 2017 review,1 John Doench discusses how the workflow of infection and subsequent cell passaging must be optimized to avoid drop out of sgRNAs, because at each passage stage only a fraction of the cells are re-seeded, the number will determine how well representation is maintained. Similarly, sampling error after the screen, during PCR from genomic DNA can result in loss of guides (and information).

For the same reason, beginning with an sgRNA library with a high degree of uniformity across the guides will minimize the likelihood that some guides will drop out before others—or conversely that some guides will be overrepresented. Uniformity is controlled at the point of manufacture and can be assessed at initial quality control and profiling of the library by next-generation sequencing.

Designing a screening assay

The basis of a CRISPR screen is the separation of cells according to the phenotype of interest, and identifying those sgRNA that have been depleted or enriched after the application of a selection pressure—for example, treatment with particular drug or small molecule inhibitor. Negative, or drop out screens, are useful when identifying oncogenes or synthetic lethal interactions of tumor cells, whereas positive screens can be utilized to identify tumor suppressors or genes involved in drug resistance. The simplest screening assay is a basic proliferation assay—but cells can also be selected by flow cytometry and, more recently, single-cell analysis. It goes without saying that optimization of the screening assay is crucial—dosages and timepoints must be selected to ensure successful gene editing but also ensure that sgRNA dropout is not due to generalized apoptosis.

Library selection—which CRISPR tool to use

The beauty of the CRISPR system is the ubiquity of possible targets and genetic perturbations—all that is required for targeting is a PAM site (Protospacer Adjacent Motif). The most commonly used Cas9 from S.pyogenes targets 5’-NGG, which occurs on average every 8bp, meaning there is an abundance of sgRNA to choose from. But not all sgRNA are created equal, as there is inherent variability in the efficiency of individual sgRNA binding, plus homology to sequences elsewhere in the genome can result in off-target activity.

Designing a library can be aided with the use of algorithms and design rules, but an easier option is to select a commercially available library. Here, the time-consuming job of library design and validation has already been done, with sgRNA selected for optimal efficiency and biological significance. As the CRISPR-Cas9 system has moved beyond gene editing, so have the availability of libraries and a range of genetic perturbations for use in screening, including CRISPR-knockout, CRISPRi, and CRISPRa. Generally, a CRISPR-ko library for a loss of function study would be a good place to start for an initial assay as they are the most developed, but once the model system and assay has been optimized it would be simple to expand screening to include the alternative Cas9 libraries.

Optimize, optimize, optimize

The recent development of CRISPR-Cas9 gene-editing technology now provides scientists with the means to perform targeted, specific manipulation of gene expression. Together with the ability to perform genome-wide CRISPR experiments simultaneously in a pooled library workflow, functional genomic screening can be performed at scale, in an unbiased manner within the technical capabilities of most laboratories. However, the simplicity and ease of access of CRISPR, and raft of CRISPR screening publications, can overshadow the significant optimization involved—which can take longer than the screen itself. But the upfront investment in perfecting each stage in the screening process will pay dividends when following up any hits that arise.

Key Takeaways

  • A CRISPR-Cas9 pooled library workflow is a multistep process, involving transduction with the lentiviral library, selection of infected cells, proliferation, seeding and passaging, isolation of genomic DNA, and PCR amplification
  • Ensuring that library representation is maintained is crucial for accurate quantification of sgRNA after screening
  • Optimization of each step in the process can take longer than the screen itself—but will pay off with increased confidence in the hits that emerge

References

1. Doench, J. G. Am I ready for CRISPR? A user’s guide to genetic screens. Nature Reviews Genetics 19, 67–80 (2018).