CRISPRpred

CRISPRpred predicts on-target activity of single guide RNAs (sgRNAs) for CRISPR/Cas9 genome editing to inform sgRNA selection and design.

Key Features:

• In Silico Prediction: Leverages a Support Vector Machine (SVM) model to predict sgRNA on-target activity from extracted sequence features.
• Performance Metrics: Evaluated on a benchmark dataset of 17 genes and 5,310 guide sequences (20% true positives) with AUROC 0.85, AUPR 0.56 (approximately 5% improvement over state-of-the-art), and maximum MCC 0.48.
• Flexibility and Feature Extraction: Extracts and integrates relevant features from sgRNA sequences to accommodate diverse genomic contexts and experimental conditions.

Scientific Applications:

• sgRNA design for CRISPR/Cas9 experiments: Guides selection of sgRNAs to improve on-target efficacy in genome editing experiments.
• Medical and genetic research: Supports sgRNA selection for medical research, therapeutic development, and genetic studies.

Methodology:

Uses a Support Vector Machine (SVM) model applied to extracted sgRNA sequence features and trains/validates performance on a comprehensive benchmark dataset of 17 genes and 5,310 guides.