Super-Charged Mass Spectral Data Processing

Automated Structural Analysis using AI and GC-HRTOFMS


Automated Structural Analysis using AI and GC-HRTOFMS

With the recent high-profile introduction of ChatGPT and Bard it’s no secret that artificial intelligence is here to stay. No matter how far this technology progresses, however, the most important software for any instrumentation will always be the scientist running the experiment. Looking to the future, AI will super-charge scientific analysis with high-speed identification of unknown compounds, and that future might be closer than we think.
At JEOL, Dr. Masaaki Ubukata’s recent work on Automated Structural Analysis for Real Unknown Compounds combines artificial intelligence (AI) with GC-HRTOFMS to rapidly provide analysis results of known mass spectra, automatically perform qualitative analysis of unknown chemical compounds, and provide high quality structural analysis results.
Prior to the development of our new software, structural analysis of real unknown compounds required significant knowledge, experience, and time: sorting through structural information for EI fragment ions, identifying higher-mass m/z ions as functional groups and substructures, looking at low-mass m/z fragment ions to classify compounds, and working through possible McLafferty rearrangement to derive structures from the fragmentation.
In this AI-assisted workflow, deconvolution of GC-MS data into mass spectrum, library search, molecular ion search, accurate mass analysis, isotope pattern matching, and structural analysis can all happen within minutes.
This newly developed automated structural analysis software, msFineAnalysis AI, two different AI programs: First, to predict chemical formulas from EI and SI (soft ionization) data and, second, to assign candidate structure matches by using our newly developed database for structure analysis.
Automated Structural Analysis using AI and GC-HRTOFMS

Soft Ionization, but Smarter

Typically, despite being the most popular method used in gas chromatography-mass spectrometry (GC-MS), electron ionization (EI) results in weak or absent molecular ions in the mass spectral data, making it difficult to identify unknowns by EI alone. However, this technology leverages soft ionization data for molecular ion information with the large number of fragment ions observed in EI mass spectra to automatically interpret this data into potential structural formulas, even if the operator has no knowledge of mass spectrometry or AI.
Because of this, the predicted structural formula of a detected component is automatically obtained, with the ability to analyze the structures of up to 50 unknown compounds in 3 minutes or less. With less time required to perform qualitative analysis for a variety of sample types containing complex mixtures of structurally-similar compounds, scientists have more free time for running additional samples or getting started on their next experiment.
msFineAnalysis AI builds upon its predecessors, msFineAnalysis and msFineAnalysis iQ, but takes the analysis a step further by providing structural formula prediction for unknown compounds not registered in the NIST database. This technology is exclusively available on the JEOL AccuTOF™ GC-Alpha, our high-resolution GC-MS with best-in-class mass resolution, accurate mass measurement, and high mass accuracy. Stop by booth 2222 or download the software brochure to learn more about how AI software can energize your analysis!



Add New Comment

© Copyright 2023 by JEOL USA, Inc.
Terms of Use
Privacy Policy
Cookie Preferences