Next-Generation CRISPR Screening Libraries Without PCR Bias and Cloning Hassle – Interview with Vivlion

(Interview is condensed and edited for clarity)

- Hi Manuel and Martin. It’s great to speak with you both today. I’d like to start with the basics. What exactly is meant by a CRISPR screen?

Martin: A CRISPR screen is a functional genomics experiment that can reveal clues about the functions of genes in a specific context. Typically, a library of plasmids, each containing one or more guide RNAs (gRNAs) targeting the investigated genes, is delivered to cells, along with a nuclease such as Cas9. Depending on the nuclease supplied, this results in a population of cells where one or more genes are edited to be knocked out, inhibited, activated, or changed in their amino acid composition. Among other uses, CRISPR screens are widely performed to investigate whether certain genes are involved in disease progression or in mediating resistance or susceptibility to certain drugs.

Manuel: The human genome encompasses about 20,000 protein-coding genes, and it’s often not clear what a given gene, or a combination of genes, does in a particular context, for instance, the contribution of a particular set of genes to a disease phenotype. Depending on how they are designed and executed, CRISPR screens can allow us, in a non-biased way, to look at and understand the function of single genes or gene combinations in a cell population or within individual cells in specific contexts or conditions.

Two main approaches to CRISPR screening

- CRISPR screens seem to be popping up in the literature on a weekly if not daily basis, and I’ve noticed much variety in the strategies and workflows being used. Can you take us through the most common approaches?

Manuel: CRISPR screening has been available for about 10 years now and it has been applied in many ways, but in essence there are two conventional approaches: CRISPR dropout screens and CRISPR enrichment screens. These forward screening approaches are independent of the biological question being asked and are widely used to identify genes that cause, for example, drug susceptibility or drug resistance in cancer cells.

In dropout screens (or negative-selection screens), cells containing gRNAs that target essential genes will be depleted from the pool of cells, while editing of genes that mediate drug resistance will increase the cells’ susceptibility to that drug. Conversely, positive-selection or enrichment screens reveal selective advantages such as drug or toxin resistance when specific genes are edited.

Curated CRISPR libraries for human and mouse whole-genome screens

- How does Vivlion choose which gRNAs go into the libraries?

Manuel: CRISPR has been accessible for roughly 12 years, and there are probably few human cell lines left that have not been subjected to a CRISPR screen. This means there is a vast amount of publicly available CRISPR screening data. We have used a combination of these data, gold-standard machine learning algorithms, and validation experiments to identify the best gRNAs, i.e., the guides that exhibit target specificity and result in the most robust and reproducible phenotypes across cell lines and experimental contexts.

Vivlion’s single PRCISR™ CRISPR libraries generally contain four gRNAs for each gene in the human or mouse genome, with minimal variance within each target-specific gRNA set. For combinatorial libraries, the number of gRNAs per gene can vary depending on the experimental scale. However, the number and design of our gRNAs makes it easier to identify genes that are truly responsible for the observed phenotype within a screen, thereby adding confidence to the data.

- What level of documentation is supplied with the PRCISR™ libraries?

Each PRCISR™ CRISPR library is subjected to rigorous QC through molecular biology techniques and next-generation sequencing (NGS). Plus, we provide analytical QC reports and NGS read-counts with each library so that the user can directly start the experiment without the need to repeat the QC step.

Experimental design and proper controls are crucial to screening success

- Besides the library challenges we have discussed, you mentioned earlier that challenges also arise within experimental design and identifying the right controls. What are the crucial considerations for experimental design?

Martin: Within the experimental design phase, the hypothesis or biological question being asked should be properly defined so that the experiment can be set up in the best possible way to answer it. For us, project success is the most important measure. That also means that we point clients to alternative solutions if we think they are superior to our services. It is also crucial to ensure that the data eventually coming out of the screen can be analysed adequately and have enough statistical power to be meaningful. For us, this means including data scientists early on in the library and screen design phase and discussing a proper control concept for the screening experiment, and, if required, also for the library.

Size and scale of the experiment should also be considered. When it is not possible to screen all the desired genes or gene combinations in a single experiment, one must downscale such that the most interesting gene combinations or sets of genes are prioritised.

- Is there a general rule of thumb when it comes to identifying controls for CRISPR screens?

Martin: Two types of controls must be included in every CRISPR screen. Firstly, one must include positive- and negative phenotype controls. These are used to confirm that the gene-editing reagents are being delivered correctly to the cells and that the editing conditions are optimally set up, and to establish a baseline of the measured phenotype at the end of the screen, respectively.

Secondly, positive editing controls are usually validated gRNAs that have demonstrated high editing efficiency across multiple cell types without inducing a phenotype. Non-targeting guides that have low sequence similarity to the genome being targeted typically serve as negative editing controls, but these should allow loading with Cas9 (or whatever nuclease is being used); this way, the negative control is mechanistically similar to the targeting and positive-control gRNAs.

Experimental controls help the researcher to minimise the influence of variables beyond the measured phenotypes. For instance, a dropout screen to detect drug sensitising genes should include untreated cells that are screened in parallel with the same library, such that conclusions can be made about the impact of gene edits and ultimately about phenotypes in the presence of the drug.

To sum up, proper preparation can result in good quality data, and errors or omissions during the design phase will likely lead to inconclusive data.

Standardisation is emerging within the CRISPR screening community

- Standard terminology, methods and tools are emerging in the genome-editing field, which is spearheaded by the NIST Genome Editing Consortium. With so many considerations and steps involved, is there a push for standardisation or benchmarking within the CRISPR screening community?

Manuel: The short answer is yes. There are four important parameters to look at when judging screen quality. These are: the number of genes being targeted, library quality, size of the experiment, and the computational methods used to analyse the data.

Pioneering research groups have performed systematic CRISPR screens in many different cell lines. Based on these and similar data, they extracted a core set of genes that result in conserved phenotypes when edited, regardless of cell type or experimental context. Among these are so-called core essential genes, and one of the QC steps in every CRISPR screening experiment should be to confirm that these essential, but also non-essential, genes perform as expected, i.e., are they really dropping out, and if so to what extent? This is a key QC step in judging the quality of a CRISPR screen.

RNA targeting screens and custom CRISPR libraries

- It sounds like the field is heading in the right direction. Before we finish up, what can we expect to see from Vivlion in the near future?

Manuel: There are a couple of developments we are fairly excited about. I’ll focus on our screening services, and I will let Martin conclude with our efforts regarding innovative reagents.

So, Cas 9 screens are great for identifying essential genes, with knockout leading to dropouts, but targeting transcripts is a really good approach to identifying essential genes that may also be drug sensitisers, since these would be ideal drug targets. So, if we expose cells to a drug at a concentration that causes a 50 % reduction in viability and perform an RNA-targeted screen at the same time, we expect to see more severe phenotypes in cells where a drug sensitiser is edited. We are very excited about what these types of projects can reveal!

We also actively explore combinatorial gene effects. Beyond identifying paralogues, these approaches provide the much needed resolution on gene synergy and can even be used to predict functionally active multi-protein complexes. Coupled to, let’s say, fluorescent reporters for specific biochemical processes, these efforts have the potential to reveal and explain highly complex biological processes and may even transform our understanding of certain biological concepts.

Martin: On the reagent side, right now, we are about to release up-to-date versions of mouse and human genome-scale libraries that are ready-to-use, and will add additional species in the very near future. In addition, we regularly supply custom libraries in various different formats and backbones, for instance, to clients who wish to use their own custom set of genes or gRNA sequences.

We also offer the full-service portfolio, meaning that we provide PRCISR™ CRISPR libraries and also perform screens including the analysis for clients. At the moment, we offer enrichment and depletion screens as Manuel said earlier, in formats of reporter screens, FACS-based enrichment screens, drug sensitiser screens, and many others.

We also help clients to select suitable algorithms for gRNA design and regularly collaborate with our clients to generate libraries for other model organisms beyond mouse and human. It may be trickier to design libraries for infrequently used model organisms, depending on how much data is already available, but we enjoy challenges and welcome informal queries from anyone who might be interested in custom libraries or screens.

Our latest product goes beyond CRISPR libraries and screens and solves the bottleneck of efficient homology-directed-repair. Here, we leverage our 3Cs expertise in generating circular-single-stranded (css) DNA to provide cssDNA to clients in sizes of up to 15-20kb. cssDNA supports highly efficient single- and multi-locus recombination without the activation of the innate immune response, which makes this type of reagent particularly suited when working with primary cells and for potential therapeutic applications. We plan to launch cssDNA for the broad market in Q2 2024 and are open for early requests.

- Thank you so much for taking the time to speak with me today. It sounds like you have some exciting times ahead!