PyGenePlexus

PyGenePlexus applies network-based supervised machine learning to predict and interpret associations between genes and input gene sets within molecular interaction networks.

Key Features:

• Molecular Interaction Networks: Uses molecular interaction networks as the foundational framework to contextualize input gene sets within broader biological systems.
• Supervised Machine Learning Model: Employs a supervised machine learning model informed by the network to predict associations between genes and the input gene set.
• Gene Association Predictions: Predicts how each gene in the network relates to the input gene set, providing candidate functional relationships and pathway associations.
• Model Interpretability via Comparative Models: Enhances interpretability by comparing the model trained on the input gene set with models trained on thousands of known gene sets to assess biological significance and novelty.
• Network Connectivity Analysis: Identifies and returns the network connectivity of the top predicted genes to reveal their potential roles and interactions within the network.

Scientific Applications:

• Gene Function Prediction: Infers candidate functions for genes based on network-contextualized machine learning associations.
• Pathway Analysis: Supports identification of pathway membership or pathway-level relationships among predicted genes.
• Biological Network Exploration: Enables exploration of molecular interaction networks to discover gene associations and network modules.
• Hypothesis Generation and Validation in Genomics and Systems Biology: Facilitates generation and comparative validation of hypotheses about gene sets in genomics and systems biology studies.

Methodology:

Train a supervised machine learning model on the input gene set within the context of a molecular interaction network and evaluate gene associations by comparing the trained model to models derived from thousands of known gene sets.