AI platform for CRISPR functional mapping and function-based drug design.

Area of Science:

• Computational biology
• Genomics
• Drug discovery

Background:

• Conventional structure-based drug design often fails clinically because binding affinity doesn't ensure functional safety.
• A gap exists between predicting drug binding and predicting safe functional modulation.

Purpose of the Study:

• To introduce CRISPRtile, a cloud-based platform for function-based drug design.
• To bridge the gap between binding affinity and safe functional modulation in drug discovery.
• To generate accurate toxicity and functional landscapes for improved drug development.

Main Methods:

• Developed library coverage optimization equations.
• Utilized AI to correct CRISPR guide biases.
• Generated toxicity and functional landscapes with reduced error compared to conventional methods.
• Enabled AI prediction of drug interaction from nontoxic functional sequences.
• Predicted brain penetration with leading performance.

Main Results:

• Achieved over threefold error reduction in toxicity and functional landscapes.
• Bypassed computational costs associated with structure-based pipelines.
• Identified FDA-approved drugs with previously unrecognized ability to modulate the NLRP3 inflammasome.
• Demonstrated AI prediction of drug interaction and brain penetration.

Conclusions:

• CRISPRtile establishes a generalizable strategy for systematic discovery of safe functional modulators.
• The platform enables function-based drug design, improving safety and efficacy.
• Identified novel modulation strategies for the NLRP3 inflammasome, impacting immune response regulation.