Knowledge-based systems, artificial neural networks and pattern recognition: Applications to biotechnological processes

Current Opinion in Biotechnology

Volumn 7, Issue 2, 1996, Pages 231-234

  Kamimura, Roy ^a   Konstantinov, Konstantin ^b   Stephanopoulos, Gregory ^a  

^a Massachusetts Institute of Technology Cambridge   (United States)

^b BAYER CORPORATION   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

ARTIFICIAL NEURAL NETWORK; BIOTECHNOLOGY; DIAGNOSIS; INDUSTRY; MONITORING; PATTERN RECOGNITION; PRIORITY JOURNAL; REVIEW;

EID: 0029863483     PISSN: 09581669     EISSN: None     Source Type: Journal    
DOI: 10.1016/S0958-1669(96)80018-8     Document Type: Article

References (18)

1. Shiba S, Nishida Y, Park Y, Iijima S, Kobayashi T: Improvement of cloned α-amylase gene expression in fed-batch culture of recombinant Saccharomyces cerevisiae by regulating both glucose and ethanol concentrations using a fuzzy controller. Biotechnol Bioeng 1994, 44:1055-1063.
2. Horiuchi H, Kamasawa M, Miyakawa H, Kishimoto M: Phase control of fed-batch culture for α-amylase production based on culture phase identification using fuzzy inference. J Ferm Bioeng 1993, 76:207-212.
3. Horiuchi J, Kamasawa M, Miyakawa H, Kishimoto M, Momose H: Effective β-galactosidase production by recombinant Escherichia coli based on culture phase identification using fuzzy set theory. J Ferm Bioeng 1993, 76-5:382-387.
4. Alfafara C, Miura K, Shimizu H, Shioya S, Suga K, Suzuki K: Fuzzy control of ethanol concentration and its application to maximum glutathione production in yeast fed-batch culture. Biotechnol Bioeng 1993, 41:493-501.
5. Kitsuta Y, Kishimoto M: Fuzzy supervisory control of glutamic acid production. Biotechnol Bioeng 1994, 44:87-94.
6. Shimizu H, Miura K, Shioya S, Suga K: On-line recognition in a yeast fed-batch culture using error vectors. Biotechnol Bioeng 1995, 47:165-173.
7. Siimes T, Linko P, Von Numers C, Nakajima M, Endo I: Real-time fuzzy-knowledge-based control of baker's yeast production. Biotechnol Bioeng 1995, 45:135-143.
8. Zhang X, Visala A, Halme A, Linko P: Functional state modeling and fuzzy control of fed-batch aerobic baker's yeast process. J Biotechnol 1994, 37:1-10.
9. Konstantinov K, Yoshida T: Physiological state control of fermentation processes. Biotechnol Bioeng 1989, 33:1145-1156.
10. Stephanopoulos G, Konstantinov K, Saner U, Yoshida T: Fermentation data analysis for diagnosis and control. In Biotechnology: a Multivolume Comprehensive Treatise, edn 2. vol 3. Edited by Rehm H-J, Reed G, Puhler A, Stadler P. Weinheim: VCH: 1995:355-400. A useful guide for determining how process variables can be used to monitor the physiological states of cells. The authors discuss how process measurements and control actions can be related to a qualitative description of the cell's physiological state. For example, gas measurements can be used for the calculation of metabolic rates. A discussion of the ways in which different control actions affect the reactor, and consequently microbial behavior, is also provided.
11. Montague G, Morris J: Neural-network contributions in biotechnology. Trends Biotechnol 1994, 12:312-324.
12. Syu M, Tsao G: Neural network modeling of batch cell growth pattern. Biotechnol Bioeng 1994, 42:376-380.
13. Glassey J, Montague G, Ward A, Kara B: Artificial neural network based experimental design procedures for enhancing fermentation development. Biotechnol Bioeng 1994, 44:397-405. Innovative use of ANNs as a method for reducing the number of experiments needed to establish operating procedures. Utilizing ANN's ability to model complex non-linear behavior, these authors train the network to predict quantities, such as biomass and protein concentration, under a wide range of operating conditions. Thus, given an appropriate training base, one could use the ANN to predict process performance in different situations without performing actual fermentation, potentially saving time and cost in process development.
14. Zhang Q, Reid J, Litchfield B, Ren J, Chang S: A prototype neural network supervised control system for Bacillus thuringiensis fermentations. Biotechnol Bioeng 1994, 43:483-489.
15. Zhang Q, Litchfield B, Reid J, Ren J, Chang S: Coupling a machine vision sensor and neural net supervised controller: controlling microbial cultivations. J Biotechnol 1995, 38:219-228.
16. Goodacre R, Trew S, Wrigley-Jones C, Neal M, Maddock J, Ottley T, Porter N, Kell D: 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.
17. Stephanopoulos G, Locher G, Duff M: Pattern recognition methods for fermentation database mining. In Preprints of the 6th International Conference on Computer Applications in Biotechnology. Edited by Munack A, Schugerl K. Garmisch-Partenkirchen: DECHEMA; 1995:195-198. Patterns can occur at different signal scales, ranging from global trends to distinct local phenomena. An example of the former might be a gradual upward increase in a particular concentration; an example of the latter could be a spike in a measurement profile. Both cases represent the extremes that patterns may take. Thus, one needs a tool that facilitates the identification of patterns at different scales. In this paper, the authors discuss how wavelets and decision trees provide a more comprehensive approach to pattern recognition.
18. Ignova, M, Glassey J, Ward A, Montague G, Irvine T: Seed analysis for production fermenter performance estimation. In Preprints of the 6th International Conference on Computer Applications in Biotechnology. Edited by Munack A, Schugerl K. Garmisch-Partenkirchen: DECHEMA; 1995:53-58. A wide variety of pattern recognition techniques (PCA, non-linear PCA, ANN, and Kohonen self-organizing feature map) are utilized to correlate final production performance to seed-tank measurements alone. Using industrial data, the authors are able to distinguish between high and low yield fermentations by assessing the quality of the 'seed' that went into the production fermenter.