Fast detection of bacterial contamination in fresh produce using FTIR and spectral classification

Screening for microbial contaminants in fresh produce is a lengthy process relative to their short shelf-life. The aim of this study is to develop a comprehensive assay which employs FTIR and spectral classification algorithm for detection of bacterial contamination of fresh produce. The procedure starts by soaking a sample of the fresh produce in broth for 5 h. Then, magnetic nanoparticles are added to capture bacteria which are then collected and prepared for FTIR scanning. The generated FTIR spectra are compared against an in-house database of different bacterial species (n = 6). The ability of the database to discriminate contaminated and uncontaminated samples and to identify the bacterial species was assessed. The compatibility of the FTIR procedures with subsequent DNA extraction and PCR was tested. The developed procedure was applied for assessment of bacterial contamination in fresh produce samples from the market (n = 78). Results were compared to the conventional culture methods. The generated FTIR database coupled to spectral classification was able to detect bacterial contamination with overall accuracy exceeding 90%. The sample processing did not alter the integrity of the bacterial DNA which was suitable for PCR. On application to fresh produce samples collected from the market, the developed method was able to detect bacterial contamination with 94% concordance with the culture method. In conclusion, the developed procedure can be applied for fast detection of microbial contamination in fresh produce with comparable accuracy to conventional microbiological assays and is compatible with subsequent molecular assays.

Decoding Optical Data with Machine Learning

Optical spectroscopy and imaging techniques play important roles in many fields such as disease diagnosis, biological study, information technology, optical science, and materials science. Over the past decade, machine learning (ML) has proved promising in decoding complex data, enabling rapid and accurate analysis of optical spectra and images. This review aims to shed light on various ML algorithms for optical data analysis with a focus on their applications in a wide range of fields. The goal of this work is to sketch the validity of ML-based optical data decoding. The review concludes with an outlook on unaddressed problems and opportunities in this emerging subject that interfaces optics, data science and ML.

Occurrence of Yeasts in White-Brined Cheeses: Methodologies for Identification, Spoilage Potential and Good Manufacturing Practices

Yeasts are generally recognized as contaminants in the production of white-brined cheeses, such as Feta and Feta-type cheeses. The most predominant yeasts species are Debaryomyces hansenii, Geotrichum candidum, Kluyveromyces marxianus, Kluyveromyces lactis, Rhodotorula mucilaginosa, and Trichosporon spp. Although their spoilage potential varies at both species and strain levels, yeasts will, in case of excessive growth, present a microbiological hazard, effecting cheese quality. To evaluate the hazard and trace routes of contamination, the exact taxonomic classification of yeasts is required. Today, identification of dairy yeasts is mainly based on DNA sequencing, various genotyping techniques, and, to some extent, advanced phenotypic identification technologies. Even though these technologies are state of the art at the scientific level, they are only hardly implemented at the industrial level. Quality defects, caused by yeasts in white-brined cheese, are mainly linked to enzymatic activities and metabolism of fermentable carbohydrates, leading to production of metabolites (CO2, fatty acids, volatile compounds, amino acids, sulfur compounds, etc.) and resulting in off-flavors, texture softening, discoloration, and swelling of cheese packages. The proliferation of spoilage yeast depends on maturation and storage conditions at each specific dairy, product characteristics, nutrients availability, and interactions with the co-existing microorganisms. To prevent and control yeast contamination, different strategies based on the principles of HACCP and Good Manufacturing Practice (GMP) have been introduced in white-brined cheese production. These strategies include milk pasteurization, refrigeration, hygienic sanitation, air filtration, as well as aseptic and modified atmosphere packaging. Though a lot of research has been dedicated to yeasts in dairy products, the role of yeast contaminants, specifically in white-brined cheeses, is still insufficiently understood. This review aims to summarize the current knowledge on the identification of contaminant yeasts in white-brined cheeses, their occurrence and spoilage potential related to different varieties of white-brined cheeses, their interactions with other microorganisms, as well as guidelines used by dairies to prevent cheese contamination.

Merging FT-IR and NGS for simultaneous phenotypic and genotypic identification of pathogenic Candida species

The rapid and accurate identification of pathogen yeast species is crucial for clinical diagnosis due to the high level of mortality and morbidity induced, even after antifungal therapy. For this purpose, new rapid, high-throughput and reliable identification methods are required. In this work we described a combined approach based on two high-throughput techniques in order to improve the identification of pathogenic yeast strains. Next Generation Sequencing (NGS) of ITS and D1/D2 LSU marker regions together with FTIR spectroscopy were applied to identify 256 strains belonging to Candida genus isolated in nosocomial environments. Multivariate data analysis (MVA) was carried out on NGS and FT-IR data-sets, separately. Strains of Candida albicans, C. parapsilosis, C. glabrata and C. tropicalis, were identified with high-throughput NGS sequencing of ITS and LSU markers and then with FTIR. Inter- and intra-species variability was investigated by consensus principal component analysis (CPCA) which combines high-dimensional data of the two complementary analytical approaches in concatenated PCA blocks normalized to the same weight. The total percentage of correct identification reached around 97.4% for C. albicans and 74% for C. parapsilosis while the other two species showed lower identification rates. Results suggested that the identification success increases with the increasing number of strains actually used in the PLS analysis. The absence of reliable FT-IR libraries in the current scenario is the major limitation in FTIR-based identification of strains, although this metabolomics fingerprint represents a valid and affordable aid to rapid and high-throughput to clinical diagnosis. According to our data, FT-IR libraries should include some tens of certified strains per species, possibly over 50, deriving from diverse sources and collected over an extensive time period. This implies a multidisciplinary effort of specialists working in strain isolation and maintenance, molecular taxonomy, FT-IR technique and chemo-metrics, data management and data basing.

Identification and differentiation of food-related bacteria: A comparison of FTIR spectroscopy and MALDI-TOF mass spectrometry

The food industry requires easy, accurate, and cost-effective techniques for microbial identification to ensure safe products and identify microbial contaminations. In this work, FTIR spectroscopy and MALDI-TOF mass spectrometry were assessed for their suitability and applicability for routine microbial diagnostics of food-related microorganisms by analyzing their robustness according to changes in incubation time and medium, identification accuracy and their ability to differentiate isolates down to the strain level. Changes in the protocol lead to a significantly impaired performance of FTIR spectroscopy, whereas they had only little effects on MALDI-TOF MS. Identification accuracy was tested using 174 food-related bacteria (93 species) from an in-house strain collection and 40 fresh isolates from routine food analyses. For MALDI-TOF MS, weaknesses in the identification of bacilli and pseudomonads were observed; FTIR spectroscopy had most difficulties in identifying pseudomonads and enterobacteria. In general, MALDI-TOF MS obtained better results (52-85% correct at species level), since the analysis of mainly ribosomal proteins is more robust and seems to be more reliable. FTIR spectroscopy suffers from the fact that it generates a whole-cell fingerprint and intraspecies diversity may lead to overlapping species borders which complicates identification. In the present study values between 56% and 67% correct species identification were obtained. On the opposite, this high sensitivity offers the opportunity of typing below the species level which was not possible using MALDI-TOF MS. Using fresh isolates from routine diagnostics, both techniques performed well with 88% (MALDI-TOF) and 75% (FTIR) correct identifications at species level, respectively.

Identification of microorganisms by FTIR spectroscopy: perspectives and limitations of the method

Fourier transform infrared (FTIR) spectroscopy was introduced in 1991 as a technique to identify and classify microbes. Since then, it has gained growing interest and has undergone a remarkable evolution. Highly sophisticated spectrometers have been developed, enabling a high sample throughput. Today, the generation of high-quality data in a short time and the application of the technique for rapid and reliable identification of microbes to the species level are well documented. What makes FTIR spectroscopy even more attractive is the fact that spectral information can also be exploited for strain typing purposes, which is particularly important for epidemiological analyses and some technological applications. Accordingly, in recent years, FTIR spectroscopy has been increasingly used for typing and classifying microorganisms below the species level. The intention of this review is to give an overview over current knowledge and strategies of using FTIR spectroscopy for species identification and to describe different approaches for strain typing.

Listeria weihenstephanensis sp. nov., isolated from the water plant Lemna trisulca taken from a freshwater pond

The phylogenetic position and phenotypic characteristics of two non-spore-forming bacilli similar to members of the genus Listeria were studied. The Gram-reaction-positive, slightly motile, facultatively anaerobic strains were isolated from the water plant Lemna trisulca sampled from a freshwater pond in Bavaria, Germany. Although no identification was possible employing the API Listeria test (bioMérieux), 16S rRNA sequence analysis confirmed a close phylogenetic similarity to Listeria rocourtiae DSM 22097T (99.0 % sequence similarity) and a more distant relationship to other Listeria species (96.0 % to Listeria monocytogenes DSM 20600T and 95.0 % similarity to Listeria grayi DSM 20601T). DNA–DNA hybridization analysis between the isolates and Listeria rocourtiae DSM 22097T yielded a similarity of 22.5 %. Analysis of partial sequences of sigB, prs, recA and HSP60 were studied and compared with those of other members of the genus Listeria and Brochothrix thermosphacta DSM 20171T supporting the relationships indicated by 16S rRNA gene sequences. The studied isolates were non-haemolytic and were not associated with cases of human or animal disease. While the results demonstrate that the strains belong to the genus Listeria , phenotypic and genotypic differences from Listeria rocourtiae DSM 22097T suggest that the strains represent a novel species for which the name Listeria weihenstephanensis sp. nov. is proposed; the type strain is WS 4560T ( = DSM 24698T = LMG 26374T), with WS 4615 ( = DSM 24699 = LMG 26375) as a second strain of the species.

Naumannella halotolerans gen. nov., sp. nov., a Gram-positive coccus of the family Propionibacteriaceae isolated from a pharmaceutical clean room and from food

Four Gram-stain-positive, aerobic bacterial strains isolated from a pharmaceutical clean room (strain WS4616T), a dessert milk product (strain WS4617) and from raw milk (strains WS4623 and WS4624) were characterized using a polyphasic approach. Phylogenetic analyses based on 16S rRNA and recA gene sequences showed that they formed a distinct lineage within the family Propionibacteriaceae . Similarity values between 16S rRNA gene sequences of the four novel strains and the type species of all genera belonging to the family Propionibacteriaceae were 89.2–94.1 %. The major cellular fatty acid was anteiso-C15 : 0 and the major polar lipids were diphosphatidylglycerol and phosphatidylglycerol. Respiratory quinones were MK-8(H4) and MK-9(H4). The cell-wall peptidoglycan of type A3γ contained ll-diaminopimelic acid, alanine, glycine and glutamic acid. The G+C content of the genomic DNA of strain WS4616T was 67.7 mol%. The whole-cell sugar pattern contained ribose, mannose, arabinose, glucose and galactose. On the basis of phenotypic and genetic data, strains WS4616T, WS4617, WS4623 and WS4624 are classified as members of a novel species in a new genus of the family Propionibacteriaceae , for which the name Naumannella halotolerans gen. nov., sp. nov. is proposed. The type strain is WS4616T ( = DSM 24323T = LMG 26184T) and three additional strains are WS4617, WS4623 and WS4624.

Micrococcus cohnii sp. nov., isolated from the air in a medical practice

Three Gram-reaction-positive bacteria, isolated from the air in a medical practice (strains WS4601(T), WS4602) or a pharmaceutical clean room (strain WS4599), were characterized using a polyphasic approach. Phylogenetic analyses based on 16S rRNA and recA gene sequences of the three novel strains showed that they formed a distinct lineage within the genus Micrococcus, sharing 16S rRNA gene sequence similarities of 96.1-98.0 % with other species of this genus. Chemotaxonomic features also supported the classification of the three novel strains within the genus Micrococcus. The major cellular fatty acids of strain WS4601(T) were anteiso-C(15 : 0) and iso-C(15 : 0), the cell-wall peptidoglycan was of type A3α (L-Lys-L-Ala), and the predominant respiratory quinones were MK-7(H(2)) and MK-8(H(2)). The polar lipid profile contained diphosphatidylglycerol and phosphatidylglycerol, but no phosphatidylinositol. The G+C content of the genomic DNA was 70.4 mol%. Numerous physiological properties were found that clearly distinguished strains WS4599, WS4601(T) and WS4602 from established members of the genus Micrococcus. Based on the phenotypic and phylogenetic data, strains WS4599, WS4601(T) and WS4602 are considered to represent three different strains of a novel species of the genus Micrococcus, for which the name Micrococcus cohnii sp. nov. is proposed. The type strain is WS4601(T) (=DSM 23974(T)=LMG 26183(T)).

Microbial biodiversity, quality and shelf life of microfiltered and pasteurized extended shelf life (ESL) milk from Germany, Austria and Switzerland

Information on factors limiting the shelf life of extended shelf life (ESL) milk produced by microfiltration and subsequent pasteurization is very limited. In this study, three different batches of ESL milk were analyzed at different stages of the production process and during storage at 4 °C, 8 °C and 10 °C in order to evaluate the changes in bacterial cell counts, microbial diversity and enzymatic quality. Additionally, detailed biodiversity analyses of 250 retail ESL milk packages produced by five manufacturers in Germany, Austria and Switzerland were performed at the end of shelf life. It was observed that microfiltration decreased the microbial loads by 5-6 log₁₀ units to lower than 1 CFU/mL. However, bacterial counts at the end of shelf life were extremely variable and ranged between <1 and 8 log₁₀ CFU/mL. 8% of all samples showed spoilage indicated by cell counts higher than 6 log₁₀ CFU/mL. The main spoilage groups of bacteria were Gram-negative post-process recontaminants (Acinetobacter, Chryseobacterium, Psychrobacter, Sphingomonas) and the spore formers Paenibacillus and Bacillus cereus, while other spore formers and Microbacterium spp. did not reach spoilage levels. Paenibacillus spp. and B. cereus apparently influenced enzymatic spoilage, as indicated by increased free fatty acid production, pH 4.6 soluble peptide fractions and off-flavors. In some cases, enzymatic spoilage was observed although microbial counts were well below 6 log₁₀ CFU/mL. Thirteen B. cereus isolates were characterized for their toxin profiles and psychrotolerance. Hbl, nhe, and cytK toxin genes were detected in ten, thirteen, and four isolates, respectively, whereas the ces gene was always absent. Interestingly, only three of the thirteen isolates could be allocated to psychrotolerant genotypes, as indicated by the major cold shock cspA gene signature. Generally, large discrepancies in microbial loads and biodiversity were observed at the end of shelf life, even among packages of the same production batch. We suggest that such unexpected differences may be due to very low cell counts after ESL treatment, causing stochastic variations of initial species distributions in individual packages. This would result in the development of significantly different bacterial populations during cold storage, including the occasional development of high numbers of pathogenic species such as B. cereus or Acinetobacter.

Species and strain identification of lactic acid bacteria using FTIR spectroscopy and artificial neural networks

Lactic acid bacteria (LAB) are a heterogeneous group of micro-organisms with high relevance for fermented food and beverages as well as human and animal health. Identification of LAB is complicated by species richness, a high degree of heterogeneity between species and the presence of very closely related species at the same time. In the present study we developed identification systems based on FTIR spectra and artificial neural networks (ANNs) for species identification of 92 LAB species out of nine genera and strain identification of probiotic and environmental Enterococcus faecium strains used as additives for animal feed. The reference databases established in this work permit reliable identification of LAB with accuracies of 93.2% at species and 97.1% at strain level for Ec. faecium, demonstrating the excellent suitability of FTIR spectroscopy as a technique applicable for simultaneous species and strain identification.