Pyrolysis mass spectrometry and its applications in biotechnology

Current Opinion in Biotechnology

Volumn 7, Issue 1, 1996, Pages 20-28

  Goodacre, Royston ^a   Kell, Douglas B ^a  

^a ABERYSTWYTH UNIVERSITY   (United Kingdom)

Author keywords

[No Author keywords available]

Indexed keywords

ANALYTIC METHOD; BACTERIUM IDENTIFICATION; BIOTECHNOLOGY; MASS SPECTROMETER; MASS SPECTROMETRY; MICROBIOLOGY; MULTIVARIATE ANALYSIS; NONHUMAN; PRIORITY JOURNAL; REVIEW;

BACTERIA (MICROORGANISMS);

EID: 0030028909     PISSN: 09581669     EISSN: None     Source Type: Journal    
DOI: 10.1016/S0958-1669(96)80090-5     Document Type: Article

References (71)

1. Porter N, Fox FM: Diversity of microbial products - discovery and application. Pesticide Sci 1993, 39:161-168.
2. 6 cultures per year.
3. Tanaka Y, Mura S: Agroactive compounds of microbial origin. Annu Rev Microbiol 1993, 47:57-87.
4. Gordon EM, Gallop MH, Campbell D, Holmes C, Bermak J, Look G, Murphy M, Needels M, Jacobs J, Sugarman J et al.: Combinatorial organic synthesis-applications to drug discovery. Eur J Med Chem 1995, 30(No SS):S337-S348. The view from Affymax, the home of combinatorial chemistry, on solid-phase synthesis. An excellent entrée to the literature.
5. Scheper T-H, Lammers F: Fermentation monitoring and process control. Curr Opin Biotechnol 1994, 5:187-191. In this review, improving the interface between the sensor and the bioprocess is highlighted as the major present hurdle to control. More on-line data, however, should lead to better process models.
6. Kell DB, Sonnleitner B: GMP - Good Modelling Practice. Trends Biotechnol 1995, 13:481-492. Many bioprocess models, particularly those of a predictive nature, do not take into account some fundamental principles, and the somewhat arcane nature of the field has allowed their proponents to get away with it. In this paper, the authors provide a guide for assisting the development of these methods and their integration into mainstream biotechnology.
7. Lee MS, Hook DJ, Kerns EH, Volk KJ, Rosenberg IE: Rapid screening of fermentation broths for flavones using tandem mass-spectrometry. Biol Mass Spectrom 1993, 22:84-88.
8. Heinzle E: Present and potential applications of mass-spectrometry for bioprocess research and control. J Biotechnol 1992, 25:81-114.
9. Hansen KF, Lauritsen FR, Degn H: An on-line sampling system for fermentation monitoring using membrane inlet massspectrometry (MIMS) - application to phenoxyacetic acid monitoring in penicillin fermentation. Biotechnol Bioeng 1994, 44:347-353.
10. Meuzelaar HLC, Haverkamp J, Hileman FD: Pyrolysis Mass Spectrometry of Recent and Fossil Biomaterials. Amsterdam: Elsevier; 1982.
11. Chapman JR: Practical Organic Mass Spectrometry. New York: Wiley & Sons; 1993.
12. Berkeley RCW, Goodacre R, Helyer RJ, Kelley T: Pyrolysis-mass spectrometry in the rapid identification of micro-organisms. Lab Practice 1990, 39:81-83.
13. Goodacre R: Characterisation and quantification of microbial systems using pyrolysis mass spectrometry: introducing neural networks to analytical pyrolysis. Microbiol Eur 1994, 2:16-22. This short review, which is aimed at the bacterial taxonomist, gives an introduction to PyMS and illustrates how the application of ANNs can benefit systematics.
14. Magee JT: Whole-organism fingerprinting. In Handbook of New Bacterial Systematics. Edited by Goodfellow M, O'Donnell AG. London: Academic Press; 1993:383-427.
15. Snyder AP, Smith PBW, Dworzanski JP, Meuzelaar HLC: Pyrolysis-gas chromatography-mass spectrometry - detection of biological warfare agents. ACS Symp Ser 1994, 541:62-84. In this paper, the US army attempts to develop analytical pyrolysis for extracting microbiological information from whole organisms in the field. The need for a simplified graphical representation, suitable for direct interpretation by non-microbiologists (as opposed to purely computerized 'black box' pattern recognition methods) is also discussed.
16. Goodfellow M: Inter-strain comparison of pathogenic micro-organisms by pyrolysis mass-spectrometry. Binary Comput Microbiol 1995, 7:54-60. The nature of PyMS as a rapid screening tool for epidemiological studies is discussed along with its lack of long-term reproducibility. Therefore, the clinician still needs to seek stable bacterial markers for national records. This paper illustrates some of the pitfalls of using ANNs simply as a black box.
17. Murdoch DA, Magee JT: A numerical taxonomic study of the Gram-positive anaerobic cocci. J Med Microbiol 1995, 43:148-155. The classifications derived from PyMS analysis compare lavourably with those established using typing methods based on conventional test reaction patterns for species of Peptostreptococcus.
18. Magee JT, Philpot C, Yang J, Hosein IK: Pyrolysis typing of isolates from a recurrence of systemic cryptococcosis. J Med Microbiol 1994, 40:165-169.
19. Freeman R, Sisson PR, Jenkins DR, Ward AC, Lightfoot NF, O'Brien SJ: Sporadic isolates of Escherichia coli O157.H7 investigated by pyrolysis mass-spectrometry. Epidemiol Infect 1995, 114:433-440. This paper illustrates how the regional Public Health Laboratory in Newcastle is exploiting PyMS. As part of an ongoing series of such studies, PyMS is once again demonstrated to be a useful addition to existing typing methods in the investigation of the epidemiology of bacteria.
20. Manchester LN, Toole A, Goodacre R: Characterisation of Carnobacterium species by pyrolysis mass spectrometry. J Appl Bacteriol 1995, 78:88-96. The typing of this group of food-spoilage organisms by PyMS compares favourably with genomic classification methods. Culture age had little effect on the PyMS spectra and short-term reproducibility over four weeks was excellent.
21. Helyer RJ, Bale SJ, Berkeley RCW: The application of pyrolysis mass-spectrometry to microbial ecology - rapid characterization of bacteria isolated from an estuarine environment J Anal Appl Pyrol 1993, 25:265-272.
22. Russell G: Pyrolysis mass-spectrometry - a fresh approach to old problems In brown algal systematics. Mar Biol 1995, 123:153-157.
23. Sanglier JJ, Whitehead D, Saddler GS, Ferguson EV, Goodfellow M: Pyrolysis mass-spectrometry as a method for the classification, identification and selection of actinomycetes. Gene 1992, 115:235-242.
24. De Nobel JG, Munnik T, Pureveen JBM, Eijkel GB, Mulder MM, Boon JJ, Van Den Ende H, Klis FM: Analysis of cell-wall mutants of Saccharomyces cerevisiae by pyrolysis mass-spectrometry. Acta Bot Neerlandica 1993, 42:505-516.
25. Halpin C, Knight ME, Foxon GA, Campbell MM, Boudet AM, Boon JJ, Chabbert B, Tollier MT, Schuch W: Manipulation of lignin quality by down-regulation of cinnamyl alcohol-dehydrogenase. Plant J 1994, 6:339-350.
26. Basile F, Voorhees KJ, Hadfield TL: Micro-organism Gram-type differentiation based on pyrolysis mass-spectrometry of bacterial fatty-acid methyl-ester extracts. Appl Environ Microbiol 1995, 61:1534-1539. PyMS of a variety of organisms reveals an interesting correlation between Gram-type and fatty acid content, Gram-positive organisms containing more branched fatty acids. A correlation is also observed between fatty acid content and pathogenicity.
27. McClennen WH, Arnold NS, Meuzelaar HLC: Field-portable hyphenated instrumentation - the birth of the tricorder. Trends Anal Chem, 1994, 13:286-293. One-dimensional methods (e.g. MS) are now being replaced by two-dimensional or higher-dimensional techniques (e.g. GC/MS), also referred to as hyphenated techniques. Miniaturization is assuming increasing importance. With due reference to Star Trek, the authors postulate that 'the next generation' of field-portable analytical instruments will be based on analytical pyrolysis.
28. Smith PBW, Snyder AP, Harden CS: Characterization of bacterial phospholipids by electrospray-ionization tandem mass-spectrometry. Anal Chem 1995, 67:1824-1830. Electrospray ionization (ESI) tandem MS is applied to the profiling of glycerophospholipids present in the chloroform/methanol extracts of four different bacterial species. Differences in the fatty acid composition for each bacterial species are readily apparent from a visual examination of the data sets.
29. Black GE, Fox A, Fox K, Snyder AP, Smith PBW: Electrospray tandem mass-spectrometry for analysis of native muramic acid in whole bacterial-cell hydrolysates. Anal Chem 1994, 66:4171-4176. Muramic acid in hydrolysates from whole microbial cells are quantified using ESI tandem MS. Compared with the more conventional GC/MS, the sample preparation is simpler and the ESt tandem MS sample analysis much more rapid (2 min compared with 30-50 mins for GC/MS); the authors suggest (correctly) that this approach could be adapted for rapid bacterial identification.
30. Snyder AP: Analyte detection In complex solid matrices with pyrolysis atmospheric-pressure chemical-ionization tandem mass-spectrometry. Trends Anal Chem, 1993, 12:296-303.
31. Smith PBW, Snyder AP: Characterization of bacteria by oxidative and non-oxidative pyrolysis-gas chromatography Ion trap mass-spectrometry. J Anal Appl Pyrol 1993, 24:199-210.
32. Lehrle R, Williams R, French C, Hammond T: Thermolysis and methanolysis of poly(beta-hydroxybutyrate) - random scission assessed by statistical-analysis of molecular-weight distributions. Macromolecules 1995, 28:4408-4414. Total thermolysis of PHB can be interpreted exclusively in terms of a random chain-scission mechanism. In this report, however, partial pyrolysis results were not consistent with random scission statistics; this implies that some kinetically favoured scissions are predominant (in fact, they occur near the ends of the molecules). The product distributions from methanolysis were also somewhat inconsistent with those expected on the basis of random scission, and the authors suggest that dimers hydrolyze more readily man do higher molecular weight species.
33. Abate R, Ballistreri A, Montaudo G, Impallomeni G: Thermal degradation of microbial poly(4-hydroxybutyrate). Macromolecules 1994, 27:332-336. The thermal degradation of poly(4-hydroxybutyrate) (P4HB) is investigated using thermogravimetry, direct PyMS and also by preparative pyrolysis and subsequent NMR analysis of the pyrolyzate. The authors' findings indicate that the thermal decomposition of P4HB yields a series of cyclic oligomers by an intramolecular exchange mechanism.
34. Lehrle RS, Williams RJ: Thermal degradation of bacterial poly(hydroxybutyric acid) - mechanisms from the dependence of pyrolysis yields on sample thickness. Macromolecules 1994, 27:3782-3789. The authors use a novel approach, employing variation in the sample thickness, that provides indirect control of the residence time of primary products in the melt and thereby facilitates the detection of secondary reactions. From quantitative measurments of 'bonus yields' and 'deficit yields' of monomeric, dimeric, trimeric, and tetrameric products, further doubt is cast on the view that the pyrolysis products from PHB can be accounted for entirely in terms of random (β-elimination) scissions.
35. Amann RI, Ludwig W, Schleifer KH: Phylogenetic identification and in-situ detection of individual microbial cells without cultivation. Microbiol Rev 1995, 59:143-169. A comprehensive account of the direct retrieval of ribosomal RNA sequences and of whole-cell oligonucleotide probing. The authors discuss how a combination of these approaches may be used both to detect specific rRNA sequences of uncultured bacteria in natural samples and to identify individual cells using microscopy or flow cytometry.
36. Kay HE, Coutinho HLC, Fattori M, Manfio GP, Goodacre R, Nut MP, Basaglia M, Beringen JE: The identification of Bradyrhizobium jeponicum strains isolated from Italian soils. Microbiology 1994, 140:2333-2339. PyMS and the RAPD method are used to characterize the survival of rhizobial strains from fields in northern Italy. PyMS (but not RAPD) showed that the derivatives from one inoculant formed two distinct populations: one like the parent strain, the other altered phenotypically. This demonstrates that PyMS can be employed to assess the release of genetically engineered microorganisms in the environment.
37. Wilkinson SR, Young M, Goodacre R, Morris JG, Farrow JAE, Collins MD: Phenotypic and genotypic differences between certain strains of Clostridium acetobutylicum. FEMS Microbiol Lett 1995, 125:199-204.
38. Aries RE, Gutteridge CS, Evans R: Rapid characterization of orange juice by pyrolysis mass-spectrometry. J Food Sci 1986, 51:1183-1186.
39. Aylott RI, Clyne AH, Fox AP, Walker DA: Analytical strategies to confirm scotch whisky authenticity. Analyst 1994, 119:1741-1746. Characteristic analytical profiles for various higher-alcohol congeners provide a valuable method for checking the authenticity of scotch whisky. PyMS allows the rapid confirmation of conclusions from chromatographic analysis; furthermore, the authors speculate that PyMS has the potential to be a useful 'stand alone' technique for authenticity analysis.
40. Goodacre R, Kell DB, Bianchi G: Neural networks and olive oil. Nature 1992, 359:594.
41. Goodacre R, Kell DB, Bianchi G: Rapid assessment of olive oil adulteration using pyrolysis mass spectrometry and artificial neural networks. J Sci Food Agric 1993, 63:297-307.
42. Galletti GC, Chiavari G, Bocchini P: Thermal-decomposition products of aspartame as determined by pyrolysis-gas chromatography mass-spectrometry. J Anal Appl Pyrol 1995, 32:137-151.
43. Beverly B, Kay PT, Voorhees KJ: Principal component analysis of the pyrolysis-mass spectra from African, Africanized hybrid, and European beeswax. J Anal Appl Pyrol 1995, 34:251-263.
44. Chiavari G, Fabbri D, Galletti GC, Mazzeo R: Use of analytical pyrolysis to characterize Egyptian painting layers. hromatographia 1995, 40:594-600. Archaeological objects are analyzed directly using PyGC/MS and Fourier-transform infrared (FT-IR) spectroscopy. PyGC/MS is shown to be a rapid approach for discriminating between wax-based binders and animal protein based binders and is able to recognize mixtures of the two products. FT-IR spectroscopy corroborates the pyrolysis findings, but sometimes provides less unequivocal results than PyGC/MS.
45. Everitt BS: Cluster Analysis. London: Edward Arnold; 1993.
46. Brown SD, Blank TB, Sum ST, Weyer LG: Chemometrics. Anal Chem 1994, 66:R315-R359. A very comprehensive review in this biennial series, discussing the application of a wide range of statistical methods within analytical chemistry.
47. Martens H, Næs T: Multivariate Calibration. New York: John Wiley & Sons; 1989.
48. Næs T, Kvaal K, Isaksson T, Miller C: Artificial neural networks In multivariate calibration. J Near Infrar Spectrosc 1993, 1:1-11.
49. Ripley BD: Neural networks and related methods for classification. J Roy Stats Soc Ser B 1994, 56:409-437. Presents a general framework for classification within which methods from statistics, ANNs, pattern recognition and machine learning can be compared. ANNs emerge as one of a class of flexible non-linear regression methods that can be used to classify via supervised learning.
50. Montague G, Morris J: Neural-network contributions in biotechnology. Trends Biotechnol 1994, 12:312-324. A review of some of the applications of a variety of ANNs methodologies to bioprocess modelling, control and pattern recognition.
51. Rumelhart DE, McClelland JL, PDP Research Group: Parallel Distributed Processing, Experiments in the Microstructure of Cognition, vols 1 and 2. Cambridge, Massachusetts: MIT Press; 1986.
52. Wasserman PD: Neural Computing: Theory and Practice. New York: Van Nostrand Reinhold; 1989.
53. Zupan J, Gasteiger J: Neural Networks for Chemists: an Introduction. Weinheim: VCH Verlagsgeesellschaft; 1993.
54. Goodacre R, Neal MJ, Kell DB: Quantitative analysis of multivariate data using artificial neural networks: a tutorial review and applications to the deconvolution of pyrolysis mass spectra Zbl Bakt 1996, in press. The implementation of ANNs for the analysis of multivariate data is reviewed, with particular reference to the analysis of pyrolysis mass spectra. The need for, and benefits of, multivariate data analysis are explained followed by a discussion of ANNs and their optimization.
55. Haykin S: Neural Networks. New York: Macmillan; 1994. A very comprehensive book telling you all you ever wanted to know about ANNs. A must for both the beginner and the more experienced neural net enthusiast
56. Dybowski R, Gant V: Artificial neural networks In pathology and medical laboratories. Lancet 1995, 346:1203-1207. A comprehensive account of the wide-ranging applications of ANNs in the clinical laboratory. ANNs are shown to extend the capabilities of the otherwise 'spacious electric filing cabinets' (computers) to the study of anatomical pathological system PAPNET (used to screen cervical smears), clinical chemistry, and general microbiology applications.
57. Sisson PR, Freeman R, Law D, Ward AC, Lightfoot NF: Rapid detection of verocytotoxin production status In Escherichia coli by artificial neural-network analysis of pyrolysis mass-spectra. J Anal Appl Pyrol 1995, 32:179-185. A combination of ANNa and PyMS is exploited to detect verocytotoxin production in E. coli isolates. Although only two exemplars (mass spectra) were used to train the 150-8-1 ANN, satisfactory generalization was (perhaps surprisingly) found.
58. Goodacre R, Hiom SJ, Cheeseman SL, Murdoch D, Weightman AJ, Wade WG: Identification and discrimination of oral asaccharolytic Eubacterium spp. using pyrolysis mass spectrometry and artificial neural networks. Curr Microbiol 1996, in press. The need for numerical methods that allow easy direct interpretation of pyrolysis mass spectra for the identification of bacteria is illustrated well with this large set of data. ANNs can be encoded simply and the results read off in a tabulated format, whereas the methods still employed by most users of PyMS involve the more complex examination of three-dimensional ordination plots and dendrograms.
59. Freeman R, Goodacre R, Sisson PR, Magee JG, Ward AC, Lightfoot NF: Rapid identification of species within the Mycobacterium tuberculosis complex by artificial neural network analysis of pyrorysis mass spectra. J Med Microbiol 1994, 40:170-173. PyMS and ANNs are used to distinguish between Mycobacterium tuberculosis and Mycobecterium bovis, irrespective of their susceptibility to antituberculosis agents. This is significant because it is not presently possible using DNA probes to differentiate between species of the M. tuberculosis complex.
60. Goodacre R, Neal MJ, Kell DB, Greenham LW, Noble WC, Harvey RG: Rapid Identification using pyrolysis mass spectrometry and artificial neural networks of Propionibacterium acnes isolated from dogs. J Appl Bacteriol 1994, 76:124-134. The classic approach for the identification of bacteria from their pyrolysis mass spectra is often subjective because it relies on the interpretation of complicated ordination plots and dendrograms; however, ANNs are a more objective alternative for this purpose. Furthermore, this paper contains the first demonstration, within microbiology, of the use of self-organizing feature maps (an unsupervised neural clustering technique) for grouping pyrolysis mass spectra.
61. Chun J, Atalan E, Ward AC, Goodfellow M: Artificial neural network analysis of pyrolysis mass-spectrometric data in the Identification of Streptomyces strains. FEMS Microbiol Lett 1993, 107:321-325.
62. Mavrovouniotis ML, Harper AM, Ifarraguerri Al: Classification of pyrolysis mass-spectra of biological materials using convex cones. J Chemometrics 1994, 8:305-331. A rather complex account of the use of convex cones to classify high-dimensional time-dependent pyrolysis mass spectra of biological samples; well worth a glance for the more mathematically inclined.
63. Goodacre R, Kell DB: Rapid and quantitative analysis of bioprocesses using pyrolysis mass spectrometry and neural networks: application to indole production. Anal Chim Acta 1993, 279:17-26.
64. Goodacre R, Edmonds AN, Kell DB: Quantitative analysis of the pyrolysls-mass spectra of complex mixtures using artificial neural networks: application to casamino acids in glycogen. J Anal Appl Pyrol 1993, 26:99-114.
65. Neal MJ, Goodacre R, Kell DB: On the analysis of pyrolysis mass spectra using artificial neural networks. Individual input scaling leads to rapid learning. In Proceedings of the World Congress on Neural Nets, San Diego. San Diego: Lawrence Erlbaum Associates Publishers; 1994:I-318-I-323. Standard back-propagation ANNs often take a long time to train. In this paper, a 100-fold decrease in the training time, with no loss in precision of prediction of the test sel, is achieved by scaling each input individually.
66. Goodacre R, Neal MJ, Kell DB: Rapid and quantitative analysis of the pyrolysis mass spectra of complex binary and tertiary mixtures using multivariate calibration and artificial neural networks. Anal Chem 1994, 66:1070-1085. A comprehensive comparison of ANNs and other supervised learning methods that employ multivariate linear regression, such as partial least squares and principal components regression, for the deconvolution of pyrolysis mass spectra obtained from binary mixtures of the protein lysozyme and nucleic acids in glycogen and tertiary mixtures of bacteria It is noteworthy that all three strains in tertiary mixtures of bacteria could be quantified accurately and simultaneously.
67. 5 in Escherichia coli. J Biotechnol 1994, 34:185-193. The combination of PyMS and ANNs is demonstrated to have potential for the screening and analysis of microbial cultures producing recombinant proteins.
68. Goodacre R, Trew S, Wrigley-Jones C, Neal MJ, Maddock J, Ottley TW, Porter N, Kell DB: Rapid screening for metabolite overproduction in fermenter broths using pyrolysis mass spectrometry with multivariate calibration and artificial neural networks. Biotechnol Bioeng 1994, 44:1205-1216. Levels of a drug of commercial interest are quantified accurately in fermentation samples of a single organism. Mutant producer strains are also analyzed, and PyMS/ANNs is demonstrated to be a rapid method for screening for the high-level production of desired substances. Furthermore, when predicting ampicillin concentration in model mixtures, ANNs are observed to be very robust to rather substantial variations in the biological background (either E. coli or S. aureus).
69. Goodacre R, Trew S, Wrigley-Jones C, Saunders G, Neal MJ, Porter N, Kell DB: Rapid and quantitative analysis of metabolites in fermenter broths using pyrolysis mass spectrometry with supervised learning: application to the screening of Penicillium chrysogenum fermentations for the overproduction of penicillins. Anal Chim Acta 1995, 313:25-43. As well as illustrating the capacity of PyMS/ANNs for the rapid screening of fermentation broths, this study compares the applicability of various squashing functions on the output nodes of otherwise identical ANNs. It is shown that ANNs employing linear output functions give more accurate estimates of determinand concentration near the edges of the concentration range than those using sigmoidal functions. In addition, these ANNs can be used successfully to extrapolate beyond the concentration range on which they have been trained.
70. Goodacre R, Kell DB: On correcting mass spectral drift using neural networks. Anal Chem 1996, in press. Reports the utility of ANNs for correcting instrumental drift. ANN or other multivariate calibration models, created using previously collected data, can be used to give accurate estimates of determinand concentration or the nature of bacteria (and indeed other materials) from newly acquired pyrolysis mass spectra by running standards each time and using ANNs to transform 'old' spectra into 'new' spectra.
71. Michie D, Spiegelhalter DJ, Taylor CC: Machine learning: neural and statistical classification. In Ellis Horwood Series in Artificial Intelligence. Edited by Campbell J. Chichester: Ellis Horwood; 1994. A useful comparison, based on a European Commission collaboration, of the main methods of artificial intelligence currently in use for classification. Which one is best depends rather on the data set used.