RESUMEN

Soft sensors based on deep learning regression models are promising approaches to predict real-time fermentation process quality measurements. However, experimental datasets are generally sparse and may contain outliers or corrupted data. This leads to insufficient model prediction performance. Therefore, datasets with a fully distributed solution space are required that enable effective exploration during model training. In this study, the robustness and predictive capability of the underlying model of a soft sensor was improved by generating synthetic datasets for training. The monitoring of intensified ethanol fermentation is used as a case study. Variational autoencoders were employed to create synthetic datasets, which were then combined with original datasets (experimental) to train neural network regression models. These models were tested on original versus augmented datasets to assess prediction improvements. Using the augmented datasets, the soft sensor predictive capability improved by 34%, and variability was reduced by 82%, based on R2 scores. The proposed method offers significant time and cost savings for dataset generation for the deep learning modeling of ethanol fermentation and can be easily adapted to other fermentation processes. This work contributes to the advancement of soft sensor technology, providing practical solutions for enhancing reliability and robustness in large-scale production.

RESUMEN

Feasible bioprocessing of lignocellulosic biomass requires the use of microbial strains with tolerance to inhibitor compounds and osmotic pressure, able to provide high product yield and productivity. In this sense, this study evaluated the potential of two non-conventional yeasts, Hansenula polymorpha CBS 4732 and Debaryomyces hansenii CBS 767, for use on biomass conversion in a biorefinery perspective. The ability of the strains to consume pentose and hexose sugars, to resist the toxic compounds present in hydrolysates, as well as to produce sugar alcohols and ethanol, was investigated. H. polymorpha showed highlighted resistance to toxic compounds and relevant ability to consume xylose and produce xylitol and ethanol under these conditions, at 37 °C. D. hansenii was a great producer of arabitol from glucose. The implications for sustainability due to the use of these yeasts in biorefineries was discussed. These results open up new perspectives for the development of the biorefinery sector.

Asunto(s)

RESUMEN

Advanced biorefineries, which aim at valorizing biomass (from agriculture, forestry, aquaculture, among others) into a wide spectrum of products and bioenergy, are seen today as key to implement a sustainable biobased economy. Although different concepts of biorefinery are currently under development, further research and improvement are still required to obtain environmentally friendly and economically feasible commercial scale biorefineries. Valorization of all biomass components and integration of different disciplines are some of the strategies that have been considered to improve the economic and environmental performance. This paper summarizes and discusses the most recent innovations and strategic orientations for the development of advanced biorefineries. Focus is given on the valorization of non-carbohydrate components of biomass (protein, acetic acid and lignin), on-site and tailor-made production of enzymes, big data analytics, and interdisciplinary efforts. The idea is to provide new insights and directions to support the development and large-scale implementation of biorefineries.

Asunto(s)

RESUMEN

L-asparaginase (ASNase) is a therapeutical enzyme used for treatment of acute lymphoblastic leukemia. ASNase products available in the market are produced by bacteria and usually present allergic response and important toxicity effects to the patients. Production of ASNase by yeasts could be an alternative to overcome these problems since yeasts have better compatibility with the human system. Recently, it was found that Leucosporidium scottii, a psychrotolerant yeast, produces ASNase. In order to advance the production of ASNase by this yeast, the present study aimed to select suitable process conditions able to maximize the production of this enzyme in a bench-scale bioreactor. Additionally, the accumulation of lipids during the enzyme production process was also determined and quantified. Experiments were carried out with the aim of selecting the most appropriate conditions of initial cell concentration (1.0, 3.5, and 5.6 g L-1), carbon source (sucrose and glycerol, individually or in mixture) and oxygen transfer rate (k L a in the range of 1.42-123 h-1) to be used on the production of ASNase by this yeast. Results revealed that the enzyme production increased when using an initial cell concentration of 5.6 g L-1, mixture of sucrose and glycerol as carbon source, and k L a of 91.72 h-1. Under these conditions, the enzyme productivity was maximized, reaching 35.11 U L-1 h-1, which is already suitable for the development of scale-up studies. Additionally, accumulation of lipids was observed in all the cultivations, corresponding to 2-7 g L-1 (32-40% of the cell dry mass), with oleic acid (C18 : 1) being the predominant compound (50.15%). Since the L-asparaginase biopharmaceuticals on the market are highly priced, the co-production of lipids as a secondary high-value product during the ASNase production, as observed in the present study, is an interesting finding that opens up perspectives to increase the economic feasibility of the enzyme production process.