Symposium Sessions (Day1, Day2, Day3)
Basic Sessions (Day1, Day2, Day3)
Young Researchers' Sessions (Day1, Day2, Day3)
Poster Presentations
- Day 2, May 16(Tue.) Room P (Foyer, Room 1004-1007)
-
2P-30 PDF
Structure Analysis of Pyrolysis Products of Polymer Samples Using Py-GC-TOFMS and Machine Learning
In gas chromatography-mass spectrometry (GC-MS), compounds can be identified by comparing electron ionization (EI) mass spectra to a database of known compounds. However, this method is limited by the EI mass spectra registered in the library, and so only compounds found in these databases can be identified. That means there are many compounds that will not only remain unidentified by conventional library searching, but may also result in false positives. We have developed a machine-learning model that can predict EI mass spectra from chemical structures. This allowed the creation of a database containing the predicted EI mass spectra of over 100 million chemical compounds found in PubChem. A library search method for increased accuracy and efficiency was also developed.