Machine learning of cellular metabolic rewiring

Abstract

Metabolic rewiring allows cells to adapt their metabolism in response to evolving environmental conditions. Traditional metabolomics techniques, whether targeted or untargeted, often struggle to interpret these adaptive shifts. Here, we introduce MetaboLiteLearner, a machine learning framework that harnesses the detailed fragmentation patterns from electron ionization (EI) collected in scan mode during gas chromatography/mass spectrometry (GC/MS) to predict abundance changes in metabolically adapted cells. When tested on breast cancer cells with different preferences to metastasize to specific organs, MetaboLiteLearner predicted the impact of metabolic rewiring on metabolites withheld from the training dataset using only the EI spectra, without metabolite identification or pre-existing knowledge of metabolic networks. The model learned captures shared and unique metabolomic shifts between brain- and lung-homing metastatic lineages, suggesting potential organ-tailored cellular adaptations. Integrating machine learning and metabolomics paves the way for new insights into complex cellular adaptations. SignificanceMetabolic rewiring--the cellular adaptation to shifts in environment and nutrients--plays key roles in many contexts, including cancer metastasis. Traditional metabolomics often falls short of capturing the nuances of these metabolic shifts. This work introduces MetaboLiteLearner, a machine learning approach that harnesses the rich fragmentation patterns from electron ionization collected in scan mode during gas chromatography/mass spectrometry, paving the way for new insights into metabolic adaptations. Demonstrating its robustness on a breast cancer model, we highlight MetaboLiteLearners potential to reshape our understanding of metabolic rewiring, with implications in diagnostics, therapeutics, and basic cell biology.