Oler, E. | Search bioRxiv

bioRxiv2025

BioTransformer4.0 a comprehensive computational tool for small molecule metabolism prediction

BioTransformer 4.0, the successor to BioTransformer 3.0, is a freely available in silico metabolism prediction tool. It integrates both knowledge-based and machine learning approaches to predict metabolites for small molecules using one of seven modules: abiotic, environmental, CYP450, phase II, enzyme commission-based, human gut microbial, and all human metabolism. It also provides a customizable sequence prediction module that allows users to simulate multi-step metabolic transformations by chaining among the first six different modules. BioTransformer 4.0 can make predictions more efficiently and accurately than the previous version, as it includes more than 130 new reaction rules, and also an optional validation module to improve the efficiency by restricting the number of predicted metabolites, due to their similarity among real human metabolites. We evaluated its performance by running the six-step all-human metabolism prediction on the DrugBank dataset of 2,457 known biotransformations, and the PhytoHub dataset of 633 known biotransformations - achieving recall values of 87.2% (resp., 91.6%) for the DrugBank (resp., PhytoHub) datasets.

bioinformatics↗

bioRxiv2024

Language model-guided anticipation and discovery of unknown metabolites

Despite decades of study, large parts of the mammalian metabolome remain unexplored. Mass spectrometry-based metabolomics routinely detects thousands of small molecule-associated peaks within human tissues and biofluids, but typically only a small fraction of these can be identified, and structure elucidation of novel metabolites remains a low-throughput endeavor. Biochemical large language models have transformed the interpretation of DNA, RNA, and protein sequences, but have not yet had a comparable impact on understanding small molecule metabolism. Here, we present an approach that leverages chemical language models to discover previously uncharacterized metabolites. We introduce DeepMet, a chemical language model that learns the latent biosynthetic logic embedded within the structures of known metabolites and exploits this understanding to anticipate the existence of as-of-yet undiscovered metabolites. Prospective chemical synthesis of metabolites predicted to exist by DeepMet directs their targeted discovery. Integrating DeepMet with tandem mass spectrometry (MS/MS) data enables automated metabolite discovery within complex tissues. We harness DeepMet to discover several dozen structurally diverse mammalian metabolites. Our work demonstrates the potential for language models to accelerate the mapping of the metabolome.

bioinformatics↗