Premium
UV /Vis spectroscopy combined with chemometrics for monitoring solid‐state fermentation with Rhizopus microsporus var. oligosporus
Author(s) -
Ito Shuri,
Barchi Augusto César,
Escaramboni Bruna,
de Oliva Neto Pedro,
Herculano Rondinelli Donizetti,
Azevedo Borges Felipe,
Romeiro Miranda Matheus Carlos,
Fernández Núñez Eutimio Gustavo
Publication year - 2017
Publication title -
journal of chemical technology and biotechnology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.64
H-Index - 117
eISSN - 1097-4660
pISSN - 0268-2575
DOI - 10.1002/jctb.5271
Subject(s) - rhizopus oligosporus , solid state fermentation , chemistry , chemometrics , food science , fermentation , bran , chromatography , bagasse , microbiology and biotechnology , biology , organic chemistry , raw material
BACKGROUND Difficulties in bioprocess monitoring are a drawback of solid‐state fermentation ( SSF ). Specifically, monitoring of enzyme activities in SSF is not an easy task. This work aimed to calibrate partial least squares ( PLS ) and artificial neural network ( ANN ) models for inferring protease and amylase activities, as well as protein concentration, from UV ‐Vis spectra of aqueous extracts of samples removed during SSF using Rhizopus microsporus var. oligosporus . RESULTS SSFs were performed using single agro‐industrial wastes (wheat bran, type II wheat flour, sugarcane bagasse and soybean meal) and ternary mixtures of them. Enzyme activities and protein concentrations in the aqueous extracts were quantified biochemically. The corresponding UV ‐Vis spectra of diluted extracts were also collected. The prediction quality of the ANN was higher than that of the PLS model. The relative errors considering the range for amylolytic and proteolytic enzymes were 4% (3–442 U g −1 ) and 6% (0–256 U g −1 ), respectively, for the best ANN architectures (8 and 6 neurons in hidden layer, respectively). CONCLUSION These results, in combination with correlation coefficients (R > 0.94), suggest that this approach is suitable for developing a chemosensor for monitoring SSFs , reducing the analytical work for quantification of enzyme activities. No satisfactory results were obtained for protein concentration. © 2017 Society of Chemical Industry