Russian traditional fermented dairy foods have been consumed for thousands of years. However, little research has focused on exploiting lactic acid bacteria (LAB) resources and analyzing the LAB composition of Russian traditional fermented dairy foods. In the present study, we cultured LAB isolated from fermented mare and cow milks, sour cream, and cheese collected from Kalmykiya, Buryats, and Tuva regions of Russia. Seven lactobacillus species and the Bifidobacterium genus were quantified by quantitative PCR. The LAB counts in these samples ranged from 3.18 to 9.77 log cfu/mL (or per gram). In total, 599 LAB strains were obtained from these samples using de Man, Rogosa, and Sharpe agar and M17 agar. The identified LAB belonged to 7 genera and 30 species by 16S rRNA and murE gene sequencing and multiplex PCR assay. The predominant LAB isolates were Lactobacillus helveticus (176 strains) and Lactobacillus plantarum (63 strains), which represented 39.9% of all isolates. The quantitative PCR results revealed that counts of 7 lactobacilli species and Bifidobacterium spp. of 30 fermented cow milk samples ranged from 1.19±0.34 (Lactobacillus helveticus in Tuva) to 8.09±0.71 (Lactobacillus acidophilus in Kalmykiya) log cfu/mL of fermented cow milk (mean ± standard error). The numbers of Bifidobacterium spp., Lb. plantarum, Lb. helveticus, and Lb. acidophilus revealed no significant difference between the 3 regions; nevertheless, Lactobacillus paracasei, Lactobacillus fermentum, Lactobacillus sakei, and Lactobacillus delbrueckii ssp. bulgaricus exhibited different degrees of variation across 3 regions. The results demonstrate that traditional fermented dairy products from different regions of Russia have complex compositions of LAB species. The diversity of LAB might be related to the type of fermented dairy product, geographical origin, and manufacturing process.

Lactobacillus delbrueckii ssp. bulgaricus (L. bulgaricus) is well known for its worldwide application in yogurt production. Flavor production and acid producing are considered as the most important characteristics for starter culture screening. To our knowledge this is the first study applying functional gene sequence multilocus sequence typing technology to predict the fermentation and flavor-producing characteristics of yogurt-producing bacteria. In the present study, phenotypic characteristics of 35 L. bulgaricus strains were quantified during the fermentation of milk to yogurt and during its subsequent storage; these included fermentation time, acidification rate, pH, titratable acidity, and flavor characteristics (acetaldehyde concentration). Furthermore, multilocus sequence typing analysis of 7 functional genes associated with fermentation time, acid production, and flavor formation was done to elucidate the phylogeny and genetic evolution of the same L. bulgaricus isolates. The results showed that strains significantly differed in fermentation time, acidification rate, and acetaldehyde production. Combining functional gene sequence analysis with phenotypic characteristics demonstrated that groups of strains established using genotype data were consistent with groups identified based on their phenotypic traits. This study has established an efficient and rapid molecular genotyping method to identify strains with good fermentation traits; this has the potential to replace time-consuming conventional methods based on direct measurement of phenotypic traits.

Lactic acid bacteria antimicrobial peptides (LABAMPs) are a class of active polypeptide produced during the metabolic process of lactic acid bacteria, which can inhibit or kill pathogenic bacteria or spoilage bacteria in food. LABAMPs have broad application in important practical fields closely related to human beings, such as food production, efficient agricultural planting, and so on. However, screening for antimicrobial peptides by biological experiment researchers is time-consuming and laborious. Therefore, it is urgent to develop a model to predict LABAMPs. In this work, we design a graph convolutional neural network framework for identifying of LABAMPs. We build heterogeneous graph based on amino acids, tripeptide and their relationships and learn weights of a graph convolutional network (GCN). Our GCN iteratively completes the learning of embedded words and sequence weights in the graph under the supervision of inputting sequence labels. We applied 10-fold cross-validation experiment to two training datasets and acquired accuracy of 0.9163 and 0.9379 respectively. They are higher that of other machine learning and GNN algorithms. In an independent test dataset, accuracy of two datasets is 0.9130 and 0.9291, which are 1.08% and 1.57% higher than the best methods of other online webservers.

In this study, the lactic acid bacterial population of Qula cheese from the Gansu and Sichuan provinces of China were isolated and identified. Eight strains of Streptococcus thermophilus were isolated, of which five strains were selected for further characterization based on their fermentation properties. The changes in a number of parameters, including titration acidity, pH, viable counts, PrtS protease activity and the production of acetaldehyde, diacetyl and organic acid, were monitored during fermentation and the storage of fermented milks produced by the respective strain. All of the strains displaying acidifying capacity and all five fermented milks maintained high viable counts of S. thermophilus from fermentation to storage. Our study found that the changes in the monitored parameters were strain-specific and varied considerably among the five tested strains. Fermented milks produced by strain IMAU80809 had the highest concentration of acetaldehyde and were most favorable in the sensory evaluation. This study confirms that Qula cheese is a good source for isolating novel lactic acid bacterial strains with different fermentation properties, which will be very useful for further development and industrialization of traditionally fermented dairy products.