GISA

GISA computes Gauss integral–based topological-geometric fingerprints of protein structures to identify and characterize knots, links, pokes, and other rare conformations.

Key Features:

• Generalized topology analysis: Applies concepts from mathematical Knot Theory without restricting analysis to predefined topologies or geometries.
• Gauss integrals as descriptors: Uses Gauss integrals as real-valued descriptors that capture local and global topological-geometric properties of protein folds.
• Multi-scale fingerprinting: Computes descriptors for entire protein chains and for all possible sub-chains to enable analysis across scales.
• Detection of rare conformations: Identifies invariant Gauss integral values that signify knots, pokes, links, knotted structures, and highly entangled conformations.
• Application to large structure sets: Demonstrated on approximately 8,000 high-resolution structures from the top8000 dataset, identifying known knots and highlighting spatial motifs in cis-trans isomerases.

Scientific Applications:

• Error detection in protein models: Identification of unusual conformations can indicate potential errors or inaccuracies in structural models.
• Protein folding and function analysis: Detection of rare structural motifs can provide insights into folding mechanisms and functional hypotheses.
• Structural bioinformatics research: Provides quantitative topological-geometric fingerprints for exploratory and targeted studies in structural bioinformatics.

Methodology:

Computes Gauss integrals for protein chains and all sub-chains to produce real-valued topological-geometric fingerprints based on Knot Theory concepts and identifies invariant values used to detect knots, links, pokes and entangled conformations, with analyses demonstrated on the top8000 (~8,000) high-resolution structure set.