Bioinformatics (Oxford, England)Jul 29, 2026

Deep learning design of guide RNA sequences without bases for precise CRISPR gene editing

CRISPR Gene Editing Weekly Brief ↗PubMed ↗DOI ↗OA ↗

Updated Aug 3, 2026

Abstract

Predicted off-target activities for both abasic and unmodified gRNAs showed strong correlation with experimental data (r ≥ 0.95).

Abasic gRNAs have been developed to enhance the specificity of CRISPR-Cas9 genome editing.
The deep neural network framework, abCRISPR, is designed for creating abasic gRNAs with reduced off-target activity.
Training with paired datasets of abasic and unmodified gRNAs helps improve predictions of off-target effects.
In vitro experiments were conducted with 97,583 mismatched substrates to evaluate off-target cleavage.
abCRISPR achieved an area under the curve (AUC) of 0.98, outperforming other deep learning methods for unmodified gRNAs.
The framework identified 58,875,004 potential CRISPR-targetable sites in the human genome with enhanced target specificity.

Simplified

CRISPR-Cas9 has become a widely used tool for genome editing. However, its off-target cleavage caused by partial sequence matches with guide RNAs (gRNAs) remains a critical limitation. Recently, abasic gRNAs (ØXØ) have been developed to enhance target specificity, but their effects vary depending on the positional sequence context. Here, we present abCRISPR, a deep neural network (DNN) framework for the rational design of ØXØ sequences with minimized off-target activity. abCRISPR leverages informative few-shot training with paired datasets of abasic and unmodified gRNAs, using high-quality random mismatch target libraries, exhaustively sequenced for mismatched off-target substrates (n = 97,583) in in vitro CRISPR-Cas9 cleavage experiments. Predicted off-target activities for both abasic and unmodified gRNAs showed strong correlation with experimental data (r ≥ 0.95, 10-fold cross-validation). Notably, these comprehensive training sets provide robust ground-truth negatives, enabling accurate and sensitive prediction of off-targets. For unmodified gRNAs, abCRISPR (AUC = 0.98) was validated to outperform existing deep learning-based methods (AUC = 0.45-0.68). When applied to the human genome, abCRISPR generated ØXØ sequences, covering 58,875,004 potent CRISPR-targetable sites with improved target specificity. Together, this work provides a comprehensive bioinformatics resource for safe and precise CRISPR-Cas9 genome editing.

Full Text

Full text is available at the source.

View Full Text ↗

Deep learning design of guide RNA sequences without bases for precise CRISPR gene editing

Abstract

Full Text

You found one interesting study. We’ll send the next 7.

What Lands in Your Inbox Each Week:

Recent issues from the crispr gene editing brief