Source: biological World (ID:ibioworld)

CRISPR gene editing technology has many applications in biomedical and other fields, from the treatment of genetic diseases, cancer, to agricultural breeding, nucleic acid detection and so on. CRISPR gene editing depends on its two components, wizard RNA (guide RNA,gRNA) is responsible for identifying and targeting target sites, and Cas enzyme is responsible for cleavage of target sites. CRISPR-Cas9 is the most widely used CRISPR system, but more and more studies have shown that it has potential risks to cut DNA directly.

In recent years, more and more CRISPR-Cas families have been discovered, among which CRISPR-Cas13 is a dazzling new star in the new CRISPR tools (especially Cas13d). Unlike Cas9, Cas13 is a targeted cutting RNA. The CRISPR system targeted by RNA brings great hope for the development of a new generation of gene editing therapy.

On July 3, 2023, researchers from New York University and Columbia University published a research paper entitled Prediction of on-target and off-target activity of CRISPR-Cas13d guide RNAs using deep learning in the journal Nature Biotechnology.

The research team combined deep learning technology with CRISPR screening to develop an artificial intelligence (AI) platform-TIGER, which can predict the on-target and off-target activity of RNA-targeted CRISPR system (CRISPR-Cas13d), and can accurately regulate the level of gene expression. This new technique paves the way for precise gene regulation in CRISPR gene editing therapy and further promotes the wide applicability of RNA-targeted CRISPR systems in human genetics and drug discovery.

Neville Sanjana, the paper's newsletter author, says that as larger data sets are collected from CRISPR filtering, the opportunity to apply complex machine learning models is getting faster and faster. With the TIGER model, we can predict the off-target activity and accurately regulate the expression level of specific genes, which makes it possible for many exciting new applications of RNA-targeted CRISPR in the biomedical field.

RNA-targeted CRISPR systems have a wide range of applications, such as RNA editing, targeted knockdown of mRNA to suppress specific gene expression, high-throughput screening of drugs, recognition of non-coding RNA, and can also be used to prevent or treat RNA virus infection.

High precision is the key to the safety of therapeutic RNA targeted CRISPR technology. In order to promote the clinical application of Cas13, we need to achieve two key goals: maximize targeting activity (on-target) and minimize miss activity (off-target). Miss activity includes mismatch (mismatches) of gRNA and target RNA, as well as insertion and deletion mutations (indels).

However, the early research on RNA targeting CRISPR system is mainly focused on targeting activity and mismatch, while the prediction of miss activity, especially insertion and deletion mutations, has not been well studied. In humans, about 1/5 of gene mutations are insertion or deletion mutations, so this is an important potential miss type that needs to be considered in CRISPR design.

In this latest paper, Neville Sanjana's team conducted a series of RNA-targeted CRISPR screening experiments in human cells and detected the activity of 200000 gRNA targeting essential genes in multiple human cell lines, including perfectly matched gRNA and off-target gRNA that led to mismatch, insertion or deletion mutations. As a result, a large Cas13d data set was generated, and the on-target and off-target activities of Cas13d gRNA were comprehensively evaluated.

The Neville Sanjana team worked with David Knowles, a machine learning expert and assistant professor of computer science at Columbia University, to train a deep learning model based on the above data and named it TIGER (Targeted Inhibition of Gene Expression via gRNA design). Comparing the results predicted by the deep learning model with the laboratory tests in human cells, TIGER can accurately predict the target activity and miss activity, which has become the first tool to predict the miss activity of the RNA targeted CIRSPR system.

David Knowles, co-author of the paper, said that with the huge data sets produced by modern high-throughput experiments, machine learning and deep learning are showing great advantages in the field of genomics. More importantly, we can also use "interpretable machine learning" to understand why the model can well predict the effect of gRNA.Previous studies by Neville Sanjana Labs have shown how to design Cas13 gRNA that can knock down a particular RNA, but now with TIGER, it is possible to further guide the design of Cas13 gRNA, striking a balance between targeted knocking and avoiding miss activity. By combining artificial intelligence (AI) with RNA-targeted CRISPR screening, the team envisages that the prediction of TIGER will help avoid unwanted miss activity and further promote the development of a new generation of RNA targeted therapy.

In this new study, the team also proved that the miss prediction of TIGER can be used to accurately regulate the level of gene expression and partially inhibit the expression of specific genes by mismatching gRNA. This is of great significance for many diseases caused by increased copy numbers of genes, such as Down syndrome, some types of schizophrenia, peroneal muscular atrophy, and some cancers caused by abnormal gene expression.

Use TIGER to design gRNA to achieve precise regulation of gene expression level.

In general, the AI prediction model developed by this study not only enhances our understanding of gRNA targeting specificity and avoiding miss, but also can achieve accurate regulation of gene expression level to some extent. This study further promotes the wide applicability of RNA-targeted CRISPR systems in human genetics and drug discovery.

Links to papers: 1. Https://www.nature.com/articles/s41587-023-01830-8