(Nanowerk News) Artificial intelligence can predict the activity of on- and off-target CRISPR tools that target RNA, not DNA, according to new research published in Natural Biotechnology (“Prediction of on-target and off-target activity of CRISPR–Cas13d guide RNAs using deep learning”).
The study, by researchers at New York University, Columbia Engineering, and the New York Genome Center, combined a deep learning model with a CRISPR display to control human gene expression in various ways—such as flicking a light switch to turn it off completely or by using a dimmer switch to turn off some activity. they. This precise gene control could be used to develop new CRISPR-based therapies.
CRISPR is a gene-editing technology with many uses in biomedicine and beyond, from treating sickle cell anemia to engineering more palatable mustard greens. These often work by targeting DNA using an enzyme called Cas9. In recent years, scientists discovered another type of CRISPR that instead targets RNA using an enzyme called Cas13.
RNA-targeting CRISPR can be used in a variety of applications, including RNA editing, knockdown of RNA to block expression of certain genes, and high-throughput screening to determine promising drug candidates. Researchers at NYU and the New York Genome Center created a platform for RNA-targeting CRISPR screens using Cas13 to better understand RNA regulation and to identify the function of non-coding RNAs. As RNA is the main genetic material in viruses including SARS-CoV-2 and the flu, RNA targeting CRISPR also holds promise for developing new methods to prevent or treat viral infections. Also, in human cells, when a gene is expressed, one of the first steps is the manufacture of RNA from the DNA in the genome.
The main objective of this study was to maximize the RNA-targeting CRISPR activity on the intended target RNA and minimize the activity on other RNAs that could have adverse side effects to cells. Off-target activity includes mismatch between guide and target RNA and insertion and deletion mutations. Previous studies of RNA-targeting CRISPR focused only on activity and target mismatch; predicting off-target activity, especially insertion and deletion mutations, has not been well studied. In the human population, about one in five mutations is an insertion or deletion, so this is an important type of potential target to consider for CRISPR design.
“Similar to DNA-targeting CRISPRs such as Cas9, we anticipate that RNA-targeting CRISPRs such as Cas13 will have an enormous impact in molecular biology and biomedical applications in the coming years,” said Neville Sanjana, professor of biology at NYU, associate professor of neuroscience and physiologist at NYU Grossman School of Medicine, core faculty member at the New York Genome Center, and co-senior author of the study. “Accurate guiding predictions and off-target identification will be of immense value to this emerging field and therapeutics.”
In their study at Natural Biotechnology, Sanjana and his colleagues performed a series of RNA-targeting CRISPR screens collected in human cells. They measured the activity of 200,000 guide RNAs that target important genes in human cells, including “perfectly matched” guide RNAs and off-target mismatches, insertions, and deletions.
Sanjana’s lab teamed up with lab machine learning expert David Knowles to engineer a deep learning model they named TIGER (Targeted Inhibition of Gene Expression via RNA guide design) that was trained on data from a CRISPR screen. Comparing predictions generated by deep learning models and laboratory tests in human cells, TIGER is able to predict on-target and off-target activity, outperforming previous models developed for on-target guidance design Cas13 and providing the first tool to predict off-target CRISPR targeting activity. RNAs.
“Machine learning and deep learning are demonstrating their power in genomics because they can leverage the large datasets that modern high-throughput experiments can now generate. Importantly, we can also use “interpretable machine learning” to understand why the model predicts that certain guides will work well,” said Knowles, assistant professor of computer science and systems biology at Columbia Engineering, a member of the core faculty at New York Genome. Center, and the study’s senior co-author.
“Our previous research shows how to design Cas13 guides that can knock down certain RNAs. With TIGER, we can now design Cas13 guides that strike a balance between on-target knockdown and avoidance of off-target activity,” said Hans-Hermann (Harm) Wessels, first co-author of the study and senior scientist at the New York Genome Center, who was previously a postdoctoral fellow in Sanjana’s laboratory.
The researchers also demonstrated that off-target TIGER predictions could be used to precisely modulate gene dose—the amount of a given gene expressed—by allowing partial inhibition of gene expression in cells with mismatched guidance. This may be useful for diseases where there are too many copies of a gene, such as Down syndrome, certain forms of schizophrenia, Charcot-Marie-Tooth disease (a hereditary nervous disorder), or in cancers where aberrant gene expression can lead to uncontrolled tumor growth.
“Our deep learning model can tell us not only how to design guide RNAs that tear down transcripts completely, but can also ‘customize’ them — for example, making them only produce 70% of the transcript of a particular gene,” said Andrew Stirn, a PhD student in Columbia Engineering and New York Genome Center, and first co-author of the study.
By combining artificial intelligence with the RNA-targeting CRISPR screen, the researchers envision that TIGER prediction will help avoid unwanted off-target CRISPR activity and further spur the development of a new generation of RNA-targeting therapies.
“As we collect data sets larger than CRISPR screens, the opportunities for implementing sophisticated machine learning models are expanding rapidly. We are fortunate to have David’s lab next door to ours to facilitate this amazing cross-disciplinary collaboration. And, with TIGER, we can predict off-target and precisely modulate gene dosage enabling many exciting new applications for RNA targeting CRISPR for biomedicine,” said Sanjana.