Three popular expression host systems Escherichia coli, Pichia pastoris and Drosophila S2 were analyzed techno-economically using HIV-1 Nef protein as the model product. On scale of 100 mg protein, the labor costs corresponded to 52-83% of the manufacturing costs. When analyzing the cost impact of the different phases (strain/cell line construction, bioreactor production, and primary purification), we found that with the microbial host systems the strain construction phase was most significant generating 56% (E. coli) and 72% (P. pastoris) of the manufacturing costs, whereas with the Drosophila S2 system the cell line construction and bioreactor production phases were equally significant (46 and 47% of the total costs, respectively). With different titers and production goal of 100 mg of Nef protein, the costs of P. pastoris and Drosophila S2 systems were about two and four times higher than the respective costs of the E. coli system. When equal titers and bioreactor working volumes (10 L) were assumed for all three systems, the manufacturing costs of the bioreactor production of the P. pastoris and Drosophila S2 systems were about two and 2.5 times higher than the respective costs of the E. coli system. V V C 2009 American Institute of Chemical Engineers Biotechnol. Prog., 25: 95-102, 2009
All rights reservedLocal calibration models can be said to be fairly commonly used within the field of near infrared (NIR) spectroscopy. In this context, the first example on the use of local calibration techniques is probably that of Davies et al. 1 Local algorithms can be beneficial by reducing the impact of a potentially nonlinear relationship between the constituent and the spectral information as well as of sample inhomogeneity. 2 However, although this has been the topic of numerous publications, there appears to be no generally accepted definition of what a local calibration technique is. From the work of Fearn and Davies 3 , the interpretation can be made that a local model is a model for which the calibration data has been selected, based on an estimate of closeness, from a larger database and that the prediction is based on the information from these selected spectra. With this interpretation, the construction of a local calibration model becomes a two-step operation, in which the first is to select the local subset of calibration data and the second is to extract a predictive model from this information. It is in these two steps that the difference between local calibration techThe performance of local calibration models for quantitative measurements of ammonium and acetate on samples from an anaerobic digestion process was examined. The local calibration methods used were locally weighted regression (LWR) and multi-layer partial least squares (ML-PLS) regression. The results of these two methods were compared to each other and to the results from the global partial least squares (PLS) model regression as well. For ammonium, both the local methods performed excellently in comparison with global PLS models. However, the results from the 150 LWR models regressed for ammonium also showed that the accuracy can be highly dependent on the different combination alternatives for model parameter settings and pre-processing alternatives. For this reason, a number of distance measures were evaluated as local subset selection methods in ML-PLS. The benefits of an optimised layer structure and the iterative approach in ML-PLS were also evaluated for ammonium. This showed that some benefits can be obtained by optimising the layer structure, at least in the sense that the number of layers can be reduced,and that there can be a significant advantage in using an iterative approach in the selection of the local subset of calibration data. The local calibration methods were also evaluated for acetate but, in this case, the benefits compared to global PLS calibration models were fairly insignificant with ML-PLS and none at all with LWR.
The glycerol and methanol concentrations in Pichia pastoris fermentations were measured on-line using Fourier transform infrared spectroscopy and an attenuated total reflection probe. Partial least squares regression was used to obtain calibration models. The models were regressed on synthetic multi-component spectra and semi-synthetic fermentation broth spectra. These were obtained by spectral addition. The accuracy for the on-line measurement of glycerol, given as standard error of prediction (SEP), was determined to 0.68 g/l, and the SEP of methanol was 0.13 g/l. We show how reliable calibration models are obtained relatively effortlessly by replacing extensive sampling from the reactor with simple mathematical manipulations of the model regression spectra.
Although near infrared (NIR) spectroscopy has been evaluated for numerous applications, the number of actual on-line or even on-site industrial applications seems to be very limited. In the present paper, the attempts to produce on-line predictions of the chemical oxygen demand (COD) in wastewater from a pulp and paper mill using NIR spectroscopy are described. The task was perceived as very challenging, but with a root mean square error of prediction of 149 mg/l, roughly corresponding to 1/10 of the studied concentration interval, this attempt was deemed as successful. This result was obtained by using partial least squares model regression, interpolated reference values for calibration purposes, and by evenly distributing the calibration data in the concentration space. This work may also represent the first industrial application of on-line COD measurements in wastewater using NIR spectroscopy.
In the production of sandpaper, a phenol formaldehyde polymer bonds the abrasive material to a backing material made of paper or a textile. The reaction takes place at elevated temperatures (90-120°C) for a long time (at least eight hours) and is thereby a very energydemanding process step. A possible future application would be to use near infrared (NIR) hyperspectral imaging (1000-2498 nm) for (1) monitoring the curing reaction and determining an endpoint for it and (2) for identifying inhomogeneous regions with low adhesion in the final product. A feasibility study was carried out on four series of resin-coated backing materials without abrasives. These were imaged at half hour intervals for eight hours of curing using a NIR line scan imager. In order to analyse the influence of the backing material on the net NIR signal from resin-coated samples, spectra of the backing material (paper, textile) were also collected using a moving grating NIR spectrometer (1100-2498 nm). Results from principal components analysis of the hyperspectral images indicated that the reaction was stabilised after five to six hours, although it continued slowly for at least 16 more hours. A relevant question was when to finish the heating (curing) and still obtain a final product of high quality. Partial least squares regression models for predicting the curing time were thus also evaluated. A calibration made on image mean spectra was used for predicting the curing time of each pixel in the full set of hyperspectral images. The predicted images showed the curing progress, inhomogeneous regions where the reaction had progressed to a slower extent and other physical abnormalities (for example air bubbles). Pixel prediction distribution analysis of the images was found useful for determining the significant number of components of the proposed regression models.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.