A machine learning approach for predicting CRISPR-Cas9 cleavage efficiencies and patterns underlying its mechanism of action
Problem
CRISPR-Cas9 sgRNAs can cleave unintended off-target genomic sites, and existing design tools poorly predict cleavage propensity.
Why it matters
Off-target cleavage can cause unintended mutations, undermining the safety of gene editing.
What the AI did
CRISTA, a machine learning algorithm, predicts the propensity of a genomic site to be cleaved by a given sgRNA, and shows bulges are not rare.
Results
Stated only as predictions more accurate than other available methodologies; no numeric metrics in abstract.
Impact
Peer-reviewed publication with 197 citations indicates research uptake of the tool.
Record
| Venue | PLoS Computational Biology |
|---|---|
| Published | 2017 |
| Work type | article |
| Authors | Shiran Abadi, Winston X. Yan, David Amar, Itay Mayrose |
| Citations | 197 (OpenAlex, at import time) |
| References | 76 |
| Model or method | CRISTA |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1371/journal.pcbi.1005807 |
| Open access | gold |
| Keywords | CRISPR-Cas9, sgRNA, off-target prediction, machine learning, CRISTA |
Importance rating
significant — thousands to millions affected
Off-target cleavage undermines safety of CRISPR gene editing, a field with broad therapeutic potential; abstract states safety relevance but no scale of impact.