DeepCRISPR: optimized CRISPR guide RNA design by deep learning
Problem
Accurately predicting sgRNA on-target knockout efficacy and off-target profile is a major challenge for effective CRISPR application.
Why it matters
Poor guide RNA design leads to ineffective or unsafe gene editing experiments.
What the AI did
DeepCRISPR unifies on-target and off-target sgRNA prediction into one deep learning framework and automates data-driven identification of sequence and epigenetic features affecting knockout efficacy.
Results
Claimed to surpass available state-of-the-art in silico tools; no numeric metrics in abstract
Impact
Provides an automated deep learning platform for optimized sgRNA design, claimed to outperform existing tools.
Record
| Venue | Genome biology |
|---|---|
| Published | 2018 |
| Work type | article |
| Authors | Guohui Chuai, Hanhui Ma, Jifang Yan, Ming Chen, Nanfang Hong, Dongyu Xue, Chi Zhou, Chenyu Zhu et al. |
| Citations | 634 (OpenAlex, at import time) |
| References | 58 |
| Model or method | DeepCRISPR (deep learning) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1186/s13059-018-1459-4 |
| Open access | gold |
| Keywords | CRISPR, sgRNA design, deep learning, off-target prediction, epigenetic features |
Importance rating
significant — thousands to millions affected
Abstract calls sgRNA efficacy/off-target prediction a major challenge for effective CRISPR application, a broad gene-editing platform; no scale or impact figures given.