Comprehensive detection of analytes in large chromatographic datasets by coupling factor analysis with a decision tree

Abstract. Environmental samples typically contain hundreds or thousands of unique organic compounds, and even minor components may provide valuable insight into their sources transformations. To understand atmospheric processes, individual are frequently identified quantified using gas chromatography–mass spectrometry. However, due to the complexity variable nature such data, data reduction is a significant bottleneck in analysis. Consequently, only subset known analytes often reported for dataset, large amounts potentially useful discarded. We present an automated approach cataloging identifying all chromatographic dataset demonstrate utility our analysis ambient aerosols. use coupled factor analysis–decision tree deconvolute peaks comprehensively catalog nearly dataset. Positive matrix factorization (PMF) small subsections multiple chromatograms applied extract factors that represent profiles mass spectra potential analytes, which detected. A decision based on peak parameters (e.g., location, width, height), relative ratios those parameters, shape, noise, retention time, spectrum discard erroneous combine determined same analyte. With approach, within section chromatogram cataloged, process repeated overlapping sections across chromatogram, generating complete list times estimated validate this compounds separation poorly resolved with similar resolution appear fraction chromatograms. As case study, method complex real-world composition particles, more than 1100 resolved, corresponding information along cataloged.