RESUMO
An automated system for the recognition of anomalies in multi-compound mixtures is described. The mixtures to be investigated are injected into a combined gas chromatograph-mass spectrometer and low resolution spectra are acquired by repetitive scanning. Using an on-line computer, all of these spectra are compared with a pre-recorded file of spectra obtained by identical analyses of a "normal" mixture. The matching procedure has been designed to allow for differences in retention times. The CASAC program calculates and plots the degree of coincidence and in this way determines whether the sample spectrum contains fewer or more fragments than the corresponding library spectrum. The system has been applied to studies on pathological urine samples.
