expansion-network

expansion-network infers gene regulatory networks by expanding Boolean network models with explicit cis-regulatory modules to integrate transcription factor binding evidence and expression data.

Key Features:

• Automated Inference: Automates inference of gene regulatory networks by using transcription factor binding evidence to specify potential regulatory connections within a Boolean network framework.
• Model Expansion: Expands Boolean network models into cis-regulatory Boolean networks by adding regulatory regions as additional nodes.
• Integration with Experimental Data: Compares the expanded model with expression data to identify regulatory functions consistent with experimental observations.
• Constraint-Based Modeling: Introduces biologically-motivated metadata to constrain models and facilitate detailed in silico analysis.
• Implementation Languages: Implemented in Python with the expansion algorithm developed in R.
• Use of SMT Solver: Employs the Z3 Satisfiability Modulo Theories (SMT) solver to handle logical constraints during model expansion, analogous to RE:IN (Yordanov et al., 2016).

Scientific Applications:

• Understanding Biological Phenomena: Models gene regulatory networks with explicit regulatory modules to study processes such as cell differentiation and oncogenesis.
• Disease Mechanism Exploration: Aids identification of causal connections between genes and transcription factors relevant to disease mechanisms.
• Data Integration: Leverages transcriptomic data to explore cis-regulatory arrangements that can produce similar transcriptomic responses.

Methodology:

Specifies possible regulatory connections from transcription factor binding evidence, expands the Boolean network by adding nodes representing regulatory regions to form a cis-regulatory Boolean network, uses the Z3 SMT solver to satisfy logical constraints, and compares the expanded model with expression data to identify regulatory functions consistent with observations.