IDSS: deformation invariant signatures for molecular shape comparison

Area of Science:

• Computational chemistry
• Structural bioinformatics
• Cheminformatics

Background:

• Existing molecular shape comparison (MSC) methods often treat flexible molecules as rigid, leading to inaccurate similarity assessments.
• Molecular flexibility and conformational changes are crucial for function but pose challenges for traditional MSC.

Purpose of the Study:

• Introduce a novel shape descriptor for flexible molecules.
• Develop a robust method for accurate molecular shape comparison of deformable structures.

Main Methods:

• Introduced the Inner Distance Shape Signature (IDSS) based on shortest paths within molecular shapes.
• Represented IDSS as a histogram, a probability distribution of inner distances between surface points.
• Reduced 3D flexible molecular shape comparison to histogram comparison.

Main Results:

• IDSS is insensitive to molecular shape deformation.
• IDSS effectively captures molecular structures, outperforming traditional shape descriptors.
• The method enables rapid comparisons, processing thousands per second.

Conclusions:

• The IDSS algorithm is robust and requires no prior knowledge of flexible regions.
• IDSS is effective in molecular search engine applications with conformational changes.
• IDSS offers a complementary tool for rigid MSC, with software available.