Evaluation of pyrolysis mass spectra may either be purely quantitative, using the spectra as fingerprints only, or may involve some degree of qualitative interpretation as to the biochemical nature, composition or structure of the sample. Although in selected cases direct visual evaluation of qualitative aspects by an experienced operator is possible, computer-assisted evaluation techniques have not only greatly refined quantitative comparisons between the spectra but are also becoming indispen-sible tools for qualitative interpretation, as shown by Burgard et al. (refs. 126, 127} in the interpretation of direct probe Py-MS spectra from nucleic acids. Computer-assisted techniques used for the evaluation of pyrolysis mass spectra include data pre-processing techniques, univariate statistical analysis techniques, multivariate statistical analysis techniques, feature selection and data reduction techniques, and display techniques. Most of these procedures are included in the special Py-MS software package developed by Eshuis et al. (ref. 45) at the F.O.M. Institute for Atomic and Molecular Physics, Amsterdam, and are available to interested scientists. Principal component, discriminant and factor analysis procedures are available in the widely used Statistical Package for the Social Sciences (SPSS) (ref. 128) and in the BMDP package (ref. 129). Other useful, widely available multivariate statistical analysis programs are the ARTHUR program package developed by Kowalski (ref. 130) and the CLUSTAN package (ref. 131). The latter programs are specifically oriented towards the cluster analysis approach in classification problems.
A full discussion of computer analysis techniques suitable for pyrolysis mass spectra is beyond the scope of this Chapter. Therefore, we shall discuss only the most essential data processing steps when applying the F.O.M program package to a selected medical problem, namely the comparative analysis of muscle biopsy samples from patients with Duchenne muscular dystrophy (DMD) as well as from non-dystrophic controls. Table 4 shows the clinical data for biopsied muscle tissue samples from three DMD patients and three non-dystrophic controls. It should be pointed out that DMD is an inherited, relentlessly progressive disease. DMD, which only manifests itself clinically in boys, is usually diagnosed at the age of four or five and leads to terminal illness at the age of thirteen or fourteen. Therefore increased involvement was noted for the 5, 7 and 8 year old boys (patients D, E and F). Preparation
Was this article helpful?