Researchers develop deep learning model for predicting CRISPR tool activity

CRISPR and AI accurately control gene expression.


CRISPR is a gene editing method with numerous applications in biomedicine and beyond, from curing sickle cell anemia to creating tastier mustard greens. It frequently operates by targeting DNA with an enzyme known as Cas9. Recently, scientists identified another form of CRISPR that uses an enzyme called Cas13 to target RNA. RNA-targeting CRISPRs have many uses, including RNA editing, knocking down RNA to limit gene expression, and high-throughput screening to identify promising drug candidates.

A deep learning algorithm and CRISPR screens are used to control how human genes are expressed in various ways, such as by switching on or off lights or adjusting the brightness of a lamp. The creation of novel CRISPR-based treatments may use these precise gene controls.

Researchers at New York University and the New York Genome Centre developed a platform for RNA-targeting CRISPR screens utilizing Cas13 to understand RNA regulation better and discover the role of non-coding RNAs. Because RNA is the primary genetic material in viruses such as SARS-CoV-2 and flu, RNA-targeting CRISPRs can potentially develop new approaches for preventing or treating viral infections. In addition, when a gene is expressed in human cells, one of the first processes is the formation of RNA from the DNA in the genome.

The new study goal is to maximize the activity of RNA-targeting CRISPRs on the chosen target RNA while minimizing activity on other RNAs that could have negative consequences on the cell. Mismatches between the guide and target RNAs and insertion and deletion mutations are examples of off-target action. 

Previously, investigations of RNA-targeting CRISPRs concentrated solely on on-target activity and mismatches; predicting off-target activity, notably insertion and deletion mutations, has received less attention. Approximately one in every five mutations is an insertion or deletion in human populations, so these are key types of possible off-targets to consider for CRISPR design.

The groundbreaking research conducted by Neville Sanjana, associate professor of biology at NYU, and the NYU team, has opened up a world of possibilities for RNA-targeting CRISPRs in biomedicine. Their latest study not only enhances our understanding of the potential applications of these CRISPRs but also paves the way for significant advancements in human genetics and drug discovery.

By developing guide RNA design rules, the team has provided a crucial framework for effectively targeting RNAs in various organisms, including notorious viruses like SARS-CoV-2. Moreover, their expertise extends to engineering protein and RNA therapeutics, offering promising avenues for novel treatments.

Additionally, they have harnessed the power of single-cell biology to uncover synergistic drug combinations that hold immense potential in combating leukemia. With this comprehensive approach and their relentless pursuit of innovation,

She said, “Similar to DNA-targeting CRISPRs such as Cas9, we anticipate that RNA-targeting CRISPRs such as Cas13 will have an outsized impact in molecular biology and biomedical applications in the coming years; accurate guide prediction and off-target identification will be of immense value for this newly developing field and therapeutics.”

The researcher and his team used a series of pooled RNA-targeting CRISPR screens in human cells to monitor the activity of 200,000 guide RNAs directed toward critical human cell genes.

They collaborated with David Knowles, a machine learning expert, to create the deep learning model TIGER (Targeted Inhibition of Gene Expression by guide RNA design), which was trained using the information from the CRISPR screens. TIGER outperformed earlier models created for Cas13 on-target guide design. He provided the first tool for predicting the off-target activity of RNA-targeting CRISPRs when comparing the predictions produced by the deep learning model and laboratory tests in human cells.

David Knowles, assistant professor of computer science and systems biology at Columbia University’s School of Engineering and Applied Science, a core faculty member at the New York Genome Center, said, “Machine learning and deep learning are showing their strength in genomics because they can take advantage of the huge datasets that can now be generated by modern high-throughput experiments. Importantly, we were also able to use “interpretable machine learning” to understand why the model predicts that a specific guide will work well.” 

The researchers also demonstrated that TIGER’s off-target predictions might be used to accurately control gene dosage in cells with mismatch guides by permitting partial suppression of gene production. This could benefit disorders with too many copies of a gene, such as Down syndrome, certain forms of schizophrenia, Charcot-Marie-Tooth disease (a hereditary nerve ailment), or tumors where abnormal gene expression can lead to uncontrolled tumor growth.

Hans-Hermann (Harm) Wessels, the study’s co-first author and a senior scientist at the New York Genome Center, who was previously a postdoctoral fellow in Sanjana’s laboratory, said, “Our earlier research demonstrated how to design Cas13 guides to knock down a particular RNA. With TIGER, we can now design Cas13 guides that strike a balance between on-target knockdown and avoiding off-target activity.”

Andrew Stirn, a Ph.D. student at Columbia Engineering and the New York Genome Center and the study’s co-first author, said, “Our deep learning model can tell us not only how to design a guide RNA that knocks down a transcript completely, but can also ‘tune’ it—for instance, having it produce only 70% of the transcript of a specific gene.”

The researchers hope that by combining artificial intelligence with an RNA-targeting CRISPR screen, TIGER’s predictions would help avoid unwanted off-target CRISPR activity and accelerate the creation of a new generation of RNA-targeting treatments.

According to researchers, The chances to use advanced machine learning models are expanding quickly as we acquire more significant datasets from CRISPR screens. We are fortunate that David’s lab is next door to ours since it has made this amazing, interdisciplinary collaboration possible. 

They can accurately control gene dosage with TIGER, which opens up a wealth of intriguing new uses for RNA-targeting CRISPRs in biomedicine. The NYU team is revolutionizing the field of biomedicine and propelling us toward a future where CRISPR-based therapies become an integral part of medical practice. 

Journal Reference:

  1. Wessels, HH., Stirn, A., Méndez-Mancilla, A. et al. Prediction of on-target and off-target activity of CRISPR–Cas13d guide RNAs using deep learning. Nature Biotechnology. DOI: 10.1038/s41587-023-01830-8
- Advertisement -

Latest Updates