Active Learning of Digenic Functions with Boolean Matrix Logic Programming

Lun Ai, Stephen Muggleton, Shi-Shun Liang, Geoff S. Baldwin

Abstract: We apply logic-based machine learning techniques to facilitate cellular engineering and drive biological discovery, using a comprehensive knowledge base of metabolic processes called a genome-scale metabolic network model (GEM). Predicted host behaviours are not always correctly described by GEMs. Learning the intricate genetic interactions within GEMs presents computational and empirical challenges. To address these difficulties, we describe a novel approach called Boolean Matrix Logic Programming (BMLP) by leveraging boolean matrices to evaluate large logic programs. We introduce a new system, BM LPactive, which efficiently explores the genomic hypothesis space by guiding informative experimentation through active learning. In contrast to subsymbolic methods, BM LPactive encodes a state-of-the-art GEM of a widely accepted bacterial host in an interpretable and logical representation using datalog logic programs. Notably, BM LPactive can successfully learn the interaction between a gene pair with 90% fewer training examples than random experimentation, overcoming the increase in experimental design space. BM LPactive enables rapid optimisation of metabolic models and offers a realistic approach to a self-driving lab for microbial engineering.