2019/12/11 by Ramtin Hosseini, Neda Hassanpour, Hosseini, Ramtin +5
Biochemistry, Genetics and Molecular Biology · Computer Science · #Bioinformatics and Genomic Networks #Computational Drug Discovery Methods #FOS: Biological sciences #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Metabolomics and Mass Spectrometry Studies #Quantitative Methods (q-bio.QM)
paper · pdf · doi:10.48550/arxiv.1912.05753
openalex publication_date 2019/12/11 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Motivation: Untargeted metabolomics comprehensively characterizes small\nmolecules and elucidates activities of biochemical pathways within a biological\nsample. Despite computational advances, interpreting collected measurements and\ndetermining their biological role remains a challenge. Results: To interpret\nmeasurements, we present an inference-based approach, termed Probabilistic\nmodeling for Untargeted Metabolomics Analysis (PUMA). Our approach captures\nmeasurements and known information about the sample under study in a generative\nmodel and uses stochastic sampling to compute posterior probability\ndistributions. PUMA predicts the likelihood of pathways being active, and then\nderives a probabilistic annotation, which assigns chemical identities to the\nmeasurements. PUMA is validated on synthetic datasets. When applied to test\ncases, the resulting pathway activities are biologically meaningful and\ndistinctly different from those obtained using statistical pathway enrichment\ntechniques. Annotation results are in agreement to those obtained using other\ntools that utilize additional information in the form of spectral signatures.\nImportantly, PUMA annotates many additional measurements.\n