GISA: using Gauss Integrals to identify rare conformations in protein structures

Area of Science:

• Structural Bioinformatics
• Biophysics
• Computational Biology

Background:

• Protein structure is crucial for function, driving research into structure determination methods.
• Existing methods often lack sensitivity to complex topological-geometric features (e.g., knots, links).
• Current approaches are typically limited to specific predefined topologies or geometries.

Purpose of the Study:

• To introduce a general method for analyzing protein topological-geometric features.
• To develop a computational tool (GISA) that generates a "fingerprint" of protein structures using knot theory descriptors.
• To enable the identification of unusual or novel protein conformations on multiple scales.

Main Methods:

• Developed the GISA algorithm to compute Gauss integrals (knot theory descriptors) for all protein sub-chains.
• Generated topological-geometric "fingerprints" for protein structures.
• Applied GISA to analyze thousands of high-resolution protein structures.

Main Results:

• GISA successfully identified known topological features like pokes and links using the basic writhe descriptor.
• Unrestricted analysis revealed previously undescribed entangled configurations and knotted conformations.
• Identified 10 known knots and 14 cis-trans isomerases with a potentially overlooked spatial motif.

Conclusions:

• GISA provides a general, scale-independent method for analyzing protein topology.
• The tool can identify known and novel complex protein conformations, aiding in fold classification and error detection.
• GISA offers significant potential benefits to the structural bioinformatics community.