AI Predicts CRISPR Tool Activity
Researchers from New York University, Columbia University and the New York Genome Center developed a platform of artificial intelligence that predicts on-and-off-target activities of CRISPR tools which target RNA rather than DNA.
The team used a CRISPR screen and a deep-learning model to control gene expression in humans. It was similar to flicking on a switch to turn them completely off or using a dimming knob to reduce their activity. The neural network that resulted, called TIGER (targeted inhibition of geneexpression via gRNA-design), was able to predict effectiveness from guide sequences and context. The team believes that the new technology will pave the path to precise gene controls in CRISPR-based treatments.
Andrew Stirn is a PhD student from Columbia Engineering at the New York Genome Center. He said that the deep learning model could tell him how to create a guide RNA which would knock down a gene completely. It can also be ‘tuned’, for example, by having it only produce 70% of the transcript. Stirn, a PhD student at Columbia Engineering and the New York Genome Center, is a co-first-author of the paper published by the researchers in Nature Biotechnology. The paper’s title, \”Predictions of on-target (and off-target) activity of CRISPR Cas13D Guide RNAs Using Deep Learning,\” was written by Stirn.
Source: