Physical and Oxidative Stability of Emulsions Stabilized with Fractionated Potato Protein Hydrolysates Obtained from Starch Production Side Stream.

This work studies the emulsifying and antioxidant properties of potato protein hydrolysates (PPHs) fractions obtained through enzymatic hydrolysis of potato protein using trypsin followed by ultrafiltration. Unfractionated (PPH1) and fractionated (PPH2 as >10 kDa, PPH3 as 10-5 kDa, PPH4 as 5-0.8 kDa, and PPH5 as <0.8 kDa) protein hydrolysates were evaluated. Pendant drop tensiometry and dilatational rheology were applied for determining the ability of PPHs to reduce interfacial tension and affect the viscoelasticity of the interfacial films at the oil-water interface. Peptides >10 kDa showed the highest ability to decrease oil-water interfacial tension. All PPH fractions predominantly provided elastic, weak, and easily stretchable interfaces. PPH2 provided a more rigid interfacial layer than the other hydrolysates. Radical scavenging and metal chelating activities of PPHs were also tested and the highest activities were provided by the unfractionated hydrolysate and the fractions with peptides >5 kDa. Furthermore, the ability of PPHs to form physically and oxidatively stable 5% fish oil-in-water emulsions (pH 7) was investigated during 8-day storage at 20 °C. Our results generally show that the fractions with peptides >5 kDa provided the highest physicochemical stability, followed by the fraction with peptides between 5 and 0.8 kDa. Lastly, promising sensory results with mostly mild attributes were obtained even at protein concentration levels that are higher than needed to obtain functional properties. The more prominent attributes (e.g., bitterness and astringency) were within an acceptable range for PPH3 and PPH4.

Assessing labelled carbon assimilation from poly butylene adipate-co-terephthalate (PBAT) monomers during thermophilic anaerobic digestion.

PBAT (poly butylene adipate-co-terephthalate) is a widely used biodegradable plastic, but the knowledge about its metabolization in anaerobic environments is very limited. In this study, the anaerobic digester sludge from a municipal wastewater treatment plant was used as inoculum to investigate the biodegradability of PBAT monomers in thermophilic conditions. The research employs a combination of 13C-labelled monomers and proteogenomics to track the labelled carbon and identify the microorganisms involved. A total of 122 labelled peptides of interest were identified for adipic acid (AA) and 1,4-butanedio (BD). Through the time-dependent isotopic enrichment and isotopic profile distributions, Bacteroides, Ichthyobacterium, and Methanosarcina were proven to be directly involved in the metabolization of at least one monomer. This study provides a first insight into the identity and genomic potential of microorganisms responsible for biodegradability of PBAT monomers during anaerobic digestion under thermophilic conditions.

Variability analysis of LC-MS experimental factors and their impact on machine learning.

BACKGROUND: Machine learning (ML) technologies, especially deep learning (DL), have gained increasing attention in predictive mass spectrometry (MS) for enhancing the data-processing pipeline from raw data analysis to end-user predictions and rescoring. ML models need large-scale datasets for training and repurposing, which can be obtained from a range of public data repositories. However, applying ML to public MS datasets on larger scales is challenging, as they vary widely in terms of data acquisition methods, biological systems, and experimental designs. RESULTS: We aim to facilitate ML efforts in MS data by conducting a systematic analysis of the potential sources of variability in public MS repositories. We also examine how these factors affect ML performance and perform a comprehensive transfer learning to evaluate the benefits of current best practice methods in the field for transfer learning. CONCLUSIONS: Our findings show significantly higher levels of homogeneity within a project than between projects, which indicates that it is important to construct datasets most closely resembling future test cases, as transferability is severely limited for unseen datasets. We also found that transfer learning, although it did increase model performance, did not increase model performance compared to a non-pretrained model.