Hybrid Neural Networks for Off-Target Prediction in Gene Editing: Integrating Sequence and Biophysical Features
Year: 2025
Type: article
Abstract: The introduction of CRISPR-Cas9 revolutionized genetics but brought challenges, notably off-target effects leading to undesired DNA changes.This research proposes a hybrid neural network integrating sequencing information, biophysical features, and CG content from gRNA sequences to predict off-target effects better.In this research, various neural network architectures, both simple and complex models with attention mechanisms, were developed and analyzed against three cell lines.Emphasis was given to F1 scores to enhance accuracy while reducing false positives.The study demonstrated that while complex models with attention mechanisms excel at identifying off-target effects, simpler architectures can also achieve excellent results for specific cell lines when paired with well-selected features.These findings will contribute to developing more reliable predictive models, improving the safety and efficacy of gene editing technologies.
DOI: 10.36838/v7i2.113
Author: Parsh Verma
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: bronze
OpenAlex ID: W7154360135