UniDesign

UniDesign models protein–nucleic acid interactions to engineer protospacer adjacent motif (PAM) recognition by Cas9 proteins and redesign PAM-interacting amino acids (PIAA) for CRISPR-Cas applications.

Key Features:

• Universal framework: Models protein–nucleic acid interactions across diverse proteins to support generalizable engineering tasks.
• CRISPR-Cas PAM engineering: Predicts how PAM-interacting amino acids (PIAA) recognize and bind preferred PAM sequences and enables modification of those interactions to relax or tighten PAM requirements.
• Proof of concept on Cas9: Applied to decode PAM–PIAA interactions across eight different Cas9 proteins and predict natural PAM sequences from native PIAA configurations with high accuracy.
• Computational redesign accuracy: Redesigns PIAA residues given natural PAM sequences, recovering native-like residues with over 70% identity and over 80% similarity.

Scientific Applications:

• Protein engineering of nucleic acid-binding proteins: Supports design and modification of proteins that interact with DNA or RNA by modeling binding preferences at the residue level.
• CRISPR-Cas specificity tuning: Enables computational customization of PAM recognition to expand or restrict Cas9 targeting ranges.
• Gene-editing research and therapeutic development: Provides computational designs for altering PAM requirements to adapt CRISPR-Cas systems for specific research or clinical contexts.

Methodology:

Computational modeling of protein–nucleic acid interactions to predict PAM recognition from PIAA configurations and computational redesign of PIAA residues, applied to analyze interactions across eight Cas9 proteins.