Quantum biological insights into CRISPR-Cas9 sgRNA efficiency from explainable-AI driven feature engineering.

Area of Science:

• Molecular Biology
• Bioinformatics
• Quantum Chemistry

Background:

• CRISPR-Cas9 technology offers powerful genetic manipulation but lacks predictive models for sgRNA efficiency.
• Current understanding of sgRNA efficiency is limited, especially across diverse organisms.

Purpose of the Study:

• To develop a novel feature set and resource for predicting sgRNA efficiency using explainable AI.
• To identify key sequence attributes for effective sgRNA design in bacteria.

Main Methods:

• Utilized an iterative Random Forest (iRF) model for feature engineering.
• Incorporated quantum chemical tensors and positional encoding for sequence attributes.
• Analyzed sgRNA efficiency in Escherichia coli and compared with Homo sapiens.

Main Results:

• Identified crucial sequence-specific traits for bacterial sgRNA design.
• Demonstrated that quantum descriptors capture complex nucleotide interactions.
• Highlighted genomic differences in CRISPR-Cas9 dynamics between E. coli and H. sapiens.

Conclusions:

• Novel encodings enhance understanding of CRISPR-Cas9 quantum biological mechanisms.
• The developed resource aids in interpreting and predicting sgRNA efficiency.
• This work provides a foundation for improved sgRNA design in various species.