A Quantum-Based Similarity Method in Virtual Screening.

Area of Science:

• Computational chemistry
• Quantum mechanics in drug discovery
• Molecular similarity searching

Background:

• Ligand-based virtual screening relies heavily on molecular similarity searching.
• Existing methods often lack the representational power to capture complex molecular relationships.
• Quantum mechanics offers a novel framework for developing advanced similarity metrics.

Purpose of the Study:

• To develop a state-of-the-art molecular similarity method inspired by quantum theory.
• To explore the use of complex numbers for molecular representation in quantum space.
• To evaluate the effectiveness of the proposed quantum-based similarity approach.

Main Methods:

• Proposed three techniques to represent molecular compounds using complex numbers.
• Developed the Standard Quantum-Based (SQB) similarity method based on complex pure Hilbert space.
• Utilized 2D fingerprints for molecular representation across MDDR, MUV, and DUD datasets.
• Performed simulated virtual screening experiments to assess method performance.

Main Results:

• The SQB method demonstrated significantly increased effectiveness in virtual screening.
• Recall of retrieved active molecules was high at the top 1% and top 5% levels.
• The enhanced representational power of molecules in complex number forms contributed to improved performance.
• SQB outperformed the Tanimoto benchmark similarity measure in simulated experiments.

Conclusions:

• Quantum mechanics provides a powerful foundation for developing advanced molecular similarity methods.
• Representing molecules in complex number formats enhances the efficacy of virtual screening.
• The Standard Quantum-Based (SQB) method shows significant promise for improving drug discovery pipelines.