EMPIRICAL PATTERN-RECOGNITION EXPERT-SYSTEM FOR MOLECULAR-WEIGHT ESTIMATION OF LOW-RESOLUTION MASS-SPECTRA

A fast, personal-computer based method of estimating molecular weights of organic compounds from low resolution mass spectra has been redesi gned and implemented with a rule-based expert system. It has a sequent ial design with a pattern recognition classifier followed by filter an d molecular weight estimator modules for each of six classes. The clas ses are nonhalobenzenes, chlorobenzenes, bromo- and bromochloroalkanes /alkenes, mono- and di-chloroalkanes/alkenes, tri-, tetra- and pentach loroalkanes/alkenes and unknowns. The classifier was derived from 106 NIST/EPA/MSDC reference spectra. The filters employ computed series of allowed molecular weights and selected base peaks for each class, exc ept unknown, to reduce misclassification. Empirical linear corrections from the training spectra are applied to two mass spectral features, MAXMASS and HIMAX1, to yield estimates and lower limits to the molecul ar weights. Extensive testing of the system was conducted with 32 test , 99 randomly chosen and 37 field gas chromatographic-mass spectrometr ic (GC-MS) spectra and results were compared to those from STIRS. The median absolute deviations from the true molecular weights of the test , random and field GC-MS spectra with the expert system were all 1 dal ton (average 5.6, 7.3, 5.9 daltons, respectively). This approach also was evaluated with 400 spectra of volatile and nonvolatile compounds o f pharmaceutical interest. The median and average absolute deviations from the true molecular weights of the 400 spectra were 2 and 10 dalto ns. Classification of the evaluation spectra, including many incomplet e spectra, was very good with accuracies of 97 (test, random and pharm aceutical) and 95% (field GC-MS).