GuidePro

GuidePro predicts and prioritizes single-guide RNAs (sgRNAs) for CRISPR/Cas9-mediated protein knockouts by integrating DNA sequence, amino acid, and protein structural information to improve sgRNA selection accuracy.

Key Features:

• Two-layer ensemble predictor: Implements a two-layer ensemble model that combines outputs from multiple predictive methods.
• sgRNA prioritization for protein knockout: Prioritizes sgRNAs specifically for CRISPR/Cas9-mediated protein-coding gene knockouts.
• Multimodal feature integration: Integrates DNA sequence features, amino acid properties, and protein structure information.
• Explicit factor modeling: Models sequence-specific sgRNA activity, frameshift probability, and characteristics of targeted amino acids.
• Ensemble combination of computational predictions: Combines diverse predictive methods and feature sets to form aggregated predictions.
• Mitigation of dataset biases: Uses multi-method integration to reduce dataset-specific biases in efficiency estimates.
• Cross-species targeting: Applies to sgRNA prioritization for protein-coding genes in human, monkey, and mouse genomes.
• Independent dataset evaluation: Demonstrated superior performance in predicting phenotypes resulting from protein loss-of-function across independent datasets.

Scientific Applications:

• CRISPR/Cas9 sgRNA selection: Prioritizes sgRNAs to improve the likelihood of effective protein knockout in gene-editing experiments.
• Prediction of loss-of-function phenotypes: Predicts phenotypic outcomes resulting from protein loss-of-function following CRISPR/Cas9 editing.
• Cross-species experimental design: Supports sgRNA selection for experiments in human, monkey, and mouse model systems.

Methodology:

Combines computational predictions in a two-layer ensemble that integrates features from DNA sequence, amino acid properties, and protein structures to model sequence-specific sgRNA activity, frameshift probability, and targeted-amino-acid effects, with evaluation on independent datasets.