CRISPR genome editing using computational approaches: A survey.

Area of Science:

• Molecular Biology
• Bioinformatics
• Genetics

Background:

• Clustered regularly interspaced short palindromic repeats (CRISPR) gene editing is a powerful tool, but its precision relies heavily on guide RNA (gRNA) design and Cas enzyme selection.
• Off-target effects remain a significant challenge, necessitating accurate prediction methods to ensure experimental reliability.

Purpose of the Study:

• To review and compare traditional and machine learning-based computational tools for designing optimal gRNA sequences.
• To evaluate methods for predicting off-target sites in CRISPR gene editing experiments.
• To assess the current state and future potential of these tools for enhancing CRISPR applications.

Main Methods:

• Literature review of existing computational tools for gRNA design and off-target prediction.
• Comparative analysis of experimental and predicting-based approaches for gRNA design.
• Evaluation of machine learning and deep learning algorithms in predicting CRISPR activity.

Main Results:

• Numerous computational tools exist for gRNA design, employing experimental and prediction-based strategies.
• Machine learning-based tools show promise for predicting on-target and off-target activities, though precision is currently limited.
• The performance of machine learning models, particularly deep learning, is dependent on training data volume and feature incorporation.

Conclusions:

• Optimizing gRNA design and Cas enzyme selection are paramount for advancing CRISPR gene editing accuracy and feasibility.
• Machine learning approaches are expected to become increasingly vital for reliable CRISPR activity prediction.
• Further development and data integration are needed to improve the precision of computational prediction models for CRISPR applications.