PyGTED

PyGTED computes the Graph Traversal Edit Distance (GTED), a dissimilarity measure between two graphs based on the minimum edit distance between strings formed from edge labels obtained through Eulerian traversals, for applications such as sequence coassembly and de novo variation detection.

Key Features:

• GTED metric: Defines graph dissimilarity as the minimum edit distance between strings derived from graph edge labels.
• Eulerian traversal string derivation: Forms strings by traversing edges via Eulerian traversals and using edge labels as characters.
• Minimum edit distance computation: Uses string edit distance as the core measure to quantify differences between traversed label sequences.
• Polynomial-time algorithm: Implements a polynomial-time algorithm specifically devised for computing GTED.
• Kernel paradigm integration: Leverages the kernel paradigm and graph kernels to decouple learning algorithms from the underlying geometric space of graphs.
• Cross-domain graph applicability: Applicable to diverse graph-based problems including genome informatics and other machine learning domains.

Scientific Applications:

• Sequence coassembly: Quantifies dissimilarity between assembly graphs to support coassembly analyses.
• De novo variation detection: Detects structural or sequence variation by comparing graph representations of genomes.
• Genome informatics: Applies GTED to graph-based genome analysis and comparison tasks.
• Protein function prediction: Supports graph-based comparisons relevant to protein function inference.
• Applied machine learning with graphs: Serves as a graph kernel-compatible distance measure for ML on graph-structured data.
• Social network and web data mining: Enables dissimilarity assessment for graph analyses in social networks and web mining.
• Security systems: Facilitates graph comparison tasks relevant to security and anomaly detection domains.

Methodology:

Generate strings from graph edge labels via Eulerian traversals, compute the minimum edit distance between those strings, and compute GTED using a polynomial-time algorithm integrated with the kernel paradigm.