Tweet
In a new breakthrough, scientists have successfully employed artificial intelligence to predict on-target and off-target activities of RNA-targeted CRISPR tools. This fascinating discovery was published in Nature Biotechnology by a cross-disciplinary team of researchers from New York University, Columbia Engineering, and the New York Genome Center.
CRISPR, a revolutionary gene editing technology, has played a pivotal role in biomedical advances, from treating conditions like sickle cell anemia to engineering better-tasting vegetables. CRISPR commonly targets DNA with the help of an enzyme called Cas9. However, in recent times, a different form of CRISPR has emerged that targets RNA via an enzyme named Cas13.
The latest research revolves around this form of RNA-targeted CRISPR and its wide array of applications such as RNA editing, blocking the expression of certain genes, and identifying promising drug candidates through high-throughput screening. Focusing on the regulation of RNA and understanding the role of non-coding RNAs, the researchers developed an innovative platform for RNA-targeted CRISPR screens using Cas13.
The team set out with an aim to maximize the effectiveness of RNA-targeted CRISPRs on intended RNA targets while minimizing undesirable off-target impacts that could lead to harmful side effects. The ability to predict off-target activity, especially insertion and deletion mutations, has been somewhat unexplored in previous research.
"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," says Neville Sanjana, a senior author of the study. He adds that identifying off-target activity and accurate guide prediction will greatly benefit this newly developing field and related therapeutics.
The research involved conducting a series of pooled RNA-targeting CRISPR screens in human cells and assessing the activity of 200,000 guide RNAs. The team then collaborated with machine learning expert David Knowles to create a deep learning model they dubbed TIGER (Targeted Inhibition of Gene Expression via guide RNA design). This model was trained on the data from the CRISPR screens and was proven capable of predicting both on-target and off-target activity.
“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,” Knowles elaborates.
Further, the researchers demonstrated TIGER's potential in precisely adjusting gene dosage by allowing partial inhibition of gene expression. This could have significant implications in tackling diseases caused by gene overexpression, including Down syndrome, certain forms of schizophrenia, and Charcot-Marie-Tooth disease, as well as in controlling tumor growth in cancers.
Excitingly, the team believes TIGER's accurate predictions can mitigate undesired off-target CRISPR activity and stimulate the development of new RNA-targeting therapies. "As we collect larger datasets from CRISPR screens, the opportunities to apply sophisticated machine learning models are growing rapidly. We are lucky to have David’s lab next door to ours to facilitate this wonderful, cross-disciplinary collaboration. And, with TIGER, we can predict off-targets and precisely modulate gene dosage which enables many exciting new applications for RNA-targeting CRISPRs for biomedicine," concludes Sanjana.
CRISPR, a revolutionary gene editing technology, has played a pivotal role in biomedical advances, from treating conditions like sickle cell anemia to engineering better-tasting vegetables. CRISPR commonly targets DNA with the help of an enzyme called Cas9. However, in recent times, a different form of CRISPR has emerged that targets RNA via an enzyme named Cas13.
The latest research revolves around this form of RNA-targeted CRISPR and its wide array of applications such as RNA editing, blocking the expression of certain genes, and identifying promising drug candidates through high-throughput screening. Focusing on the regulation of RNA and understanding the role of non-coding RNAs, the researchers developed an innovative platform for RNA-targeted CRISPR screens using Cas13.
The team set out with an aim to maximize the effectiveness of RNA-targeted CRISPRs on intended RNA targets while minimizing undesirable off-target impacts that could lead to harmful side effects. The ability to predict off-target activity, especially insertion and deletion mutations, has been somewhat unexplored in previous research.
"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," says Neville Sanjana, a senior author of the study. He adds that identifying off-target activity and accurate guide prediction will greatly benefit this newly developing field and related therapeutics.
The research involved conducting a series of pooled RNA-targeting CRISPR screens in human cells and assessing the activity of 200,000 guide RNAs. The team then collaborated with machine learning expert David Knowles to create a deep learning model they dubbed TIGER (Targeted Inhibition of Gene Expression via guide RNA design). This model was trained on the data from the CRISPR screens and was proven capable of predicting both on-target and off-target activity.
“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,” Knowles elaborates.
Further, the researchers demonstrated TIGER's potential in precisely adjusting gene dosage by allowing partial inhibition of gene expression. This could have significant implications in tackling diseases caused by gene overexpression, including Down syndrome, certain forms of schizophrenia, and Charcot-Marie-Tooth disease, as well as in controlling tumor growth in cancers.
Excitingly, the team believes TIGER's accurate predictions can mitigate undesired off-target CRISPR activity and stimulate the development of new RNA-targeting therapies. "As we collect larger datasets from CRISPR screens, the opportunities to apply sophisticated machine learning models are growing rapidly. We are lucky to have David’s lab next door to ours to facilitate this wonderful, cross-disciplinary collaboration. And, with TIGER, we can predict off-targets and precisely modulate gene dosage which enables many exciting new applications for RNA-targeting CRISPRs for biomedicine," concludes Sanjana.