Boolean-network simplification and rule fitting to unravel chemotherapy resistance in non-small cell lung cancer

2025-06-03

AI Summary

This study addresses the challenge of deciphering chemotherapy resistance mechanisms in non-small cell lung cancer (NSCLC). We propose the first lossless Boolean network reduction method, compressing the original 31-node resistance network into a minimal 9-node core model while fully preserving all three clinically relevant steady states and their basins of attraction—achieving over a 10⁴-fold reduction in state space. Our approach integrates synchronous and asynchronous dynamical simulations, logic-rule fitting, and a novel network reduction algorithm to ensure exact recapitulation of the original attractor landscape and resistance frequency distribution, with strong concordance to clinical data. The resulting core model substantially enhances computational efficiency in identifying control nodes and designing therapeutic interventions. It provides an interpretable, computationally tractable theoretical framework for discovering resistance-targeting biomarkers and optimizing combination therapies.

Technology Category

• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Search and Optimization: Combinatorial Optimization
• Machine Learning: Learning on the Edge & Model Compression

Application Category

• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Boolean networks are powerful frameworks for capturing the logic of gene-regulatory circuits, yet their combinatorial explosion hampers exhaustive analyses. Here, we present a systematic reduction of a 31-node Boolean model that describes cisplatin- and pemetrexed-resistance in non-small-cell lung cancer to a compact 9-node core that exactly reproduces the original attractor landscape. The streamlined network shrinks the state space by four orders of magnitude, enabling rapid exploration of critical control points, rules fitting and candidate therapeutic targets. Extensive synchronous and asynchronous simulations confirm that the three clinically relevant steady states and their basins of attraction are conserved and reflect resistance frequencies close to those reported in clinical studies. The reduced model provides an accessible scaffold for future mechanistic and drug-discovery studies.

Problem

• Simplify Boolean networks to analyze chemotherapy resistance mechanisms
• Reduce model complexity while preserving original attractor landscape
• Identify therapeutic targets for non-small-cell lung cancer

Innovation

• Reduced 31-node Boolean model to 9-node core
• Shrunk state space by four orders of magnitude
• Conserved clinically relevant steady states