Open Access

BACKGROUND Various neutral and alkaline peptidases are commercially available for use in protein hydrolysis under neutral to alkaline conditions. However, the hydrolysis of proteins under acidic conditions by applying fungal aspartic peptidases (FAPs) has not been investigated in depth so far. The aim of this study, thus, was to purify a FAP from the commercial enzyme preparation, ROHALASE® BXL, determine its biochemical characteristics, and investigate its application for the hydrolysis of food and animal feed proteins under acidic conditions. RESULTS A Trichoderma reesei derived FAP, with an apparent molecular mass of 45.8 kDa (sodium dodecyl sulfate–polyacrylamide gel electrophoresis; SDS-PAGE) was purified 13.8-fold with a yield of 37% from ROHALASE® BXL. The FAP was identified as an aspartate protease (UniProt ID: G0R8T0) by inhibition and nano-LC-ESI-MS/MS studies. The FAP showed the highest activity at 50°C and pH 4.0. Monovalent cations, organic solvents, and reducing agents were tolerated well by the FAP. The FAP underwent an apparent competitive product inhibition by soy protein hydrolysate and whey protein hydrolysate with apparent Ki-values of 1.75 and 30.2 mg*mL−1, respectively. The FAP showed promising results in food (soy protein isolate and whey protein isolate) and animal feed protein hydrolyses. For the latter, an increase in the soluble protein content of 109% was noted after 30 min. CONCLUSION Our results demonstrate the applicability of fungal aspartic endopeptidases in the food and animal feed industry. Efficient protein hydrolysis of industrially relevant substrates such as acidic whey or animal feed proteins could be conducted by applying fungal aspartic peptidases. © 2022 Society of Chemical Industry.

Micronization of biochar (BC) may ease its application in agriculture. For example, fine biochar powders can be applied as suspensions via drip-irrigation systems or can be used to produce grnulated fertilizers. However, micronization may effect important physical biochar properties like the water holding capacity (WHC) or the porosity.

In dem Projekt BioMeth wurde der Ansatz der Membranbegasung zur Erhöhung der Verfügbarkeit von gelöstem Wasserstoff für die biologische Methanisierung im Sinn der Etablierung eines Power-to-Gas-Konzeptes zur Energiespeicherung verfolgt. Übergeordnetes Ziel war die Entwicklung eines skalierbaren Verfahrenskonzeptes, dass sich zur Nutzung CO2-haltiger Gasvolumenström eignet. Geplant war es, das Verfahren am Beispiel der Biogasanlage der Biokäserei Monte-Ziego in Teningen zu demonstrieren und dort das bestehende Konzept der parallelen Abwasseraufbereitung und Energieerzeugung zu erweitern. Die ursprüngliche Struktur des Arbeitspaketplanes ist in nachfolgender Abbildung gezeigt.

Dort, wo Modelle der operativen Energiesystemanalyse untereinander Überschneidungen aufweisen, stellt sich zunächst die Frage, ob sie bei gleichgearteten Fragestellungen auch die gleichen Antworten liefern. Dies zu beantworten war erstes Ziel des hier beschriebenen Vorhabens. Das zweite Ziel war, im Falle von Differenzen zu ermitteln, worin diese begründet liegen. Es waren nicht nur die Modelle selbst, sondern auch das methodische Vorgehen zur Modellerstellung und Simulation in Betracht zu ziehen. Die darauf aufbauende Identifikation von individuellen Optimierungspotenzialen war das dritte Ziel. Da die operative Energiesystemanalyse noch ein recht junger Forschungsbereich ist, existiert darüber hinaus Klärungsbedarf, welches Modell sich für welche Untersuchungen besonders eignet und welches methodische Vorgehen sich empfiehlt. Die Beantwortung dieser Fragen stellte das vierte Ziel des Vorhabens dar.

Convolutional neural networks (CNN) define the state-of-the-art solution on many perceptual tasks. However, current CNN approaches largely remain vulnerable against adversarial perturbations of the input that have been crafted specifically to fool the system while being quasi-imperceptible to the human eye. In recent years, various approaches have been proposed to defend CNNs against such attacks, for example by model hardening or by adding explicit defence mechanisms. Thereby, a small “detector” is included in the network and trained on the binary classification task of distinguishing genuine data from data containing adversarial perturbations. In this work, we propose a simple and light-weight detector, which leverages recent findings on the relation between networks’ local intrinsic dimensionality (LID) and adversarial attacks. Based on a re-interpretation of the LID measure and several simple adaptations, we surpass the state-of-the-art on adversarial detection by a significant margin and reach almost perfect results in terms of F1-score for several networks and datasets. Sources available at: https://github.com/adverML/multiLID