LogicNet

LogicNet reconstructs gene regulatory networks and identifies logical interactions among regulatory genes using a probabilistic continuous logic framework applied directly to continuous gene expression data.

Key Features:

• Probabilistic continuous logic framework: Implements probabilistic continuous logics to model regulatory relationships without discretizing expression levels.
• Direct operation on continuous data: Operates on continuous gene expression data without thresholding or converting to Boolean or multi-state representations.
• Simultaneous structure and interaction inference: Reconstructs GRN structure while identifying logical interactions among regulatory genes in a single framework.
• No prior network required: Infers gene-gene interactions without requiring an a priori known network structure.
• Logic function detection: Identifies logical functions among regulatory genes and detects logic functions from known regulatory gene-target interactions.
• Comparative performance: Has been shown to outperform traditional fuzzy logics and common approaches such as mutual information-based and regression-based methods in accuracy and biological relevance.

Scientific Applications:

• Gene regulatory network reconstruction: Reconstruction of GRNs from continuous gene expression data.
• Logic function discovery: Detection of logical functions in known regulatory gene-target interactions.
• Exploration of unknown regulatory architectures: Analysis of complex biological systems where the underlying regulatory architecture is unknown or incomplete.

Methodology:

Applies a probabilistic continuous logic framework directly to continuous gene expression data without discretization to simultaneously reconstruct GRN structures and identify logical interactions, enabling inference of gene-gene interactions and detection of logic functions from regulatory gene-target data.