pmotif_lib

Perform motif detection, either with traditional frequency, or with positional metrics for each graphlet occurrence.

This library implements each step of a (p)motif detection pipeline such as

• graphlet detection in networks
• randomization of networks
• frequency comparison of graphlet occurrences
• positional metric comparison of graphlet occurrences.

Setup

Install this package:

pip install pmotif-lib

This library relies on the gtrieScanner tool. Please download it, compile it, and add the executable to your path.

Finally, this library loads environment variables. Create an .env file::

export DATASET_DIRECTORY=/path/where/edgelists/are/located
export EXPERIMENT_OUT=/path/where/raw/graphlets/and/pmetric/should/go
export GTRIESCANNER_EXECUTABLE=/path/to/the/gtriescanner/executable
export WORKERS=1  # Optional, controls the max. degree of parallelization

In order to make these variablea available to the library, use source .env before command line usage! You can also use the python-dotenv package to load the .env file from python code:

from dotenv import load_dotenv

load_dotenv(".env")

Usage

showcase/ contains a number of examples:

• graphlet detection
• p-graphlet detection
• motif detection
• p-motif detection

After installing pmotif_lib, navigate into showcase/, run source .showcase_env, and then run python3 graphlet_detection (or one of the other examples). This expects the gtrieScanner executable to be in your system's path under gtrieScanner!

Glossary

• Induced Subgraph: A graph created by cutting out a set of nodes from a graph G, retaining all edges between these nodes
• Isomorphic graphs: Graphs, that are structurally the same when ignoring node labels
• Isomorphic Classes of Size k: A set of graphs with k nodes, so that every other graph with k nodes is isomorphic to one graph in the set
• k-Graphlet: An isomorphic class of size k, so that at least one induced subgraph in a graph G is isomorphic to that class
• Graphlet Occurrences: All induced sub-graphs in a graph G that belong to a specific k-Graphlet
• Graphlet Frequency: The number of graphlet occurrences of a specific k-graphlet in a graph G
• Graph Motif: A k-Graphlet, which has a graphlet frequency which is significantly higher than expected, usually tested against randomized graphs generated based on G
• p-Motif: A k-Graphlet, which has graphlet occurrences with a significant expression of a positional metric, when compared against randomized graphs generated based on G