• info@icbcongress.com
• English

ساخت اکانت کاربری رایگان

• ورود

INTERNATIONAL CONGRESS OF

PharmacoGenetics

• Artificial Intelligence Integration with CRISPR/Cas9: Toward Precision Genome Editing

• Helia Salmasi,^1,* sara ghasemi ,² asiyeh jebelli,³
  1. Department of Cell and Molecular Biology, Faculty of Biological Sciences, Kharazmi University, Tehran, Iran
  2. Department of Cell and Molecular Biology, Faculty of Biological Sciences, Kharazmi University, Tehran, Iran
  3. Department of Cell and Molecular Biology, Faculty of Biological Sciences, Kharazmi University, Tehran, Iran
  
  
  
• Introduction: Clustered regularly interspaced short palindromic repeats (CRISPR) stores fragments of bacteriophage DNA from past infections, enabling recognition and destruction of similar viruses in future invasions. This antiviral defense mechanism of bacteria was used to design CRISPR/Cas9, a groundbreaking genetic engineering system. Acting as molecular scissors, the CRISPR/Cas9 system employs a guide RNA (gRNA) to direct the Cas9 nuclease to a precise genomic location, where it cleaves double-stranded DNA and activates the cell’s natural repair mechanisms. Scientists using the CRISPR/Cas9 system can manipulate to add, remove, or modify specific DNA segments with exceptional precision. This powerful technology drives transformative applications in research, medicine, and biotechnology, from foundational studies to potential treatments for genetic disorders. The integration of artificial intelligence (AI) and deep learning with CRISPR/Cas9 is revolutionizing genetic engineering and personalized medicine. Advanced AI algorithms automate critical aspects of CRISPR, such as predicting gRNA activity, optimizing Cas9 variants, and identifying anti-CRISPR proteins. Deep learning models accurately assess Cas9 activity and specificity, enabling the selection of the most effective enzymes. By analyzing extensive datasets, including genetic profiles and health histories, AI predicts treatment responses and identifies optimal genomic targets, minimizing off-target effects and improving the accuracy of CRISPR interventions.
• Methods: An extensive literature search of PubMed and Google Scholar was conducted using the terms 'CRISPR', 'artificial intelligence', and 'machine learning', together with their conceptual interconnections, to retrieve a comprehensive, focused set of relevant studies. Titles and abstracts were initially screened, followed by full-text assessment and critical evaluation of the shortlisted publications. Four articles were ultimately selected as the primary references for this research.
• Results: The convergence of machine learning (ML) with CRISPR/Cas9 technology has established a new paradigm in genome engineering, where computational frameworks accelerate discovery and reshape experimental strategies. Traditional rule-based and empirical approaches often fail to capture the complex interplay of factors influencing CRISPR efficiency and specificity, such as sequence context, chromatin accessibility, protein–DNA interactions, and experimental conditions. Recent advances in protein structure prediction tools, like trRosettaX-single and CeI-TASSER, have enabled modeling of Cas proteins without reliance on homologous templates, expanding the ability to identify mutational hotspots and rationally design Cas derivatives with enhanced functionality. These models provide insights into unresolved protein conformations and hypothesize how structural variability influences editing outcomes. A key challenge in CRISPR/Cas9 gene editing is predicting and mitigating off-target effects, where gRNAs bind unintended genomic loci, causing unwanted mutations and genomic instability. To address this, various ML approaches have been employed, including Artificial Neural Networks (ANNs), Support Vector Machines (SVMs), k-Nearest Neighbors (K-NN), Naïve Bayes, and Logistic Regression. Each model has distinct strengths and limitations. In contrast, ensemble-based methods like the Extra Trees Classifier (ETC) demonstrate superior performance, offering robust handling of complex feature interactions, effective noise reduction, and resilience to data imbalance with minimal parameter tuning, making them ideal for reliable gRNA off-target site prediction. The mutational space of Cas proteins, often exceeding one thousand amino acids, poses a significant challenge for conventional rational design and random mutagenesis. To address this, ML-assisted directed evolution frameworks have been developed, relying on iterative test–learn–design cycles. In these cycles, variant libraries are experimentally screened (test), ML models infer sequence–function landscapes (learn), and novel variants are generated based on predictive outcomes (design). Such strategies have reduced screening burden by up to 80% in some Cas9 engineering contexts while maintaining enrichment of high-performing variants. Supervised ML algorithms also predict gRNA efficiency and off-target activity by learning feature–outcome relationships from large datasets, reducing the need for exhaustive empirical testing. Explainable ML approaches (logistic regression and tree-based models) enable ranking of feature importance, guiding hypothesis generation and experimental prioritization. This transparency enhances model interpretability and bridges computational predictions with biological mechanism discovery. Furthermore, integrating epigenomic, transcriptomic, and structural datasets into ML frameworks extends predictive power across diverse organisms and cellular contexts.
• Conclusion: These innovations underscore the transformative role of ML in CRISPR/Cas9 engineering. By enabling accurate structure prediction, efficient mutational exploration, and interpretable modeling of editing outcomes, ML provides a scalable, data-driven framework for protein engineering and genome editing optimization. These advancements accelerate the development of next-generation Cas variants and deepen our understanding of CRISPR biology, establishing ML as a cornerstone of genome engineering evolution.
• Keywords: CRISPR, artificial intelligence, machine learning