Modelling growth of Bacillus cereus in paneer by one-step parameter estimation.

Bacillus cereus phylogenetic group III and IV strains are commonly associated with food products and cause toxin mediated foodborne diseases. These pathogenic strains have been identified from milk and dairy products, such as reconstituted infant formula and several cheeses. Paneer is a fresh, soft cheese originating from India that is prone to foodborne pathogen contamination, such as by Bacillus cereus. However, there are no reported studies of B. cereus toxin formation in paneer or predictive models quantifying growth of the pathogen in paneer under different environmental conditions. This study assessed enterotoxin-producing potential of B. cereus group III and IV strains, isolated from dairy farm environments, in fresh paneer. Growth of a four-strain cocktail of toxin-producing B. cereus strains was measured in freshly prepared paneer incubated at 5-55 °C and modelled using a one-step parameter estimation combined with bootstrap re-sampling to generate confidence intervals for model parameters. The pathogen grew in paneer between 10 and 50 °C and the developed model fit the observed data well (R2 = 0.972, RMSE = 0.321 log10 CFU/g). The cardinal parameters for B. cereus growth in paneer along with the 95% confidence intervals were: µopt 0.812 log10 CFU/g/h (0.742, 0.917); Topt is 44.177 °C (43.16, 45.49); Tmin is 4.405 °C (3.973, 4.829); Tmax is 50.676 °C (50.367, 51.144). The model developed can be used in food safety management plans and risk assessments to improve safety of paneer while also adding to limited information on B. cereus growth kinetics in dairy products.

Quantitative modeling of the survival of Listeria monocytogenes in soy sauce-based acidified food products.

Primary and secondary models were developed for quantitatively characterizing the survival of Listeria monocytogenes in soy-sauce based acidified Asian style products that do not undergo a thermal treatment. The objective of this study was to quantify the effect of food matrix properties on L. monocytogenes' survival in soy sauce-based products. This quantification enables a product-specific estimation of 5-log reduction time to ensure a safe processing and management operation, to ultimately facilitate a science-based, safety-oriented product development process. A central composite design with four independent variables (pH, soy sauce, added NaCl and soluble solids) with five levels was used to plan the challenge studies on different formulations. To model microbial survival over time, different non-linear primary models were fit to the data obtained from challenge studies. The best-fit model was selected based on a series of statistical goodness-of-fit measures. Kinetic parameters estimated from the best-fit primary models were fit to response surface equations using second order polynomial regression. The best-fit primary model representative of the product formulations was a modified Weibull model. The natural logarithm of the scale parameter (δ, in h) was used as the response variable for the secondary model. This resulted in acceptable fitting compared to the observed values with R2 values of 0.95 and RMSE of 0.7 h. External validity of model predictions was conducted by comparing them to 5-log reduction times observed in independent challenge tests using different product formulations. Results indicated an acceptable validation with R2 = 0.81 and RMSE = 35 h. The present study provides quantitative tools specific for cold-fill-hold soy sauce-based products to enhance microbial safety management plans and product development.

Detection, quantification, and characterization of Salmonella enterica in mango, tomato, and raw chicken purchased in the central region of Mexico.

To estimate human exposure to Salmonella enterica, it is essential to understand the pathogen distribution and characteristics. Prevalence and concentration of S. enterica were determined in mango, tomato, and raw chicken samples purchased in three states (Aguascalientes, Querétaro, and Guadalajara) located in the central region of Mexico during two seasons. In addition, S. enterica isolates were characterized by absence/presence of 13 virulence genes (chromosomal, prophage, and plasmid) and resistance to 14 antibiotics. A total of 300 samples of mango, 272 of tomato, and 354 of raw chicken were analyzed. The mean of the prevalence (24.9%) and concentration (-0.61 Log MPN/g) of S. enterica in chicken was higher than in mango (1.3%, -1.7 Log MPN/g) and tomato (1.1%, -1.7 Log MPN). Among S. enterica isolates (284), there were 7 different virulotypes, belonging 68.7% of isolates to V2; there was high variability in the presence of mobile genetic elements. The occurrence of specific mobile elements ranged from 81.4% to 11.3% among isolates. Among the isolates, 91.5% were resistant to at least one antibiotic with ampicillin being the most frequent; 54.9% of isolates were multidrug resistant. Data from this study can be used for quantitative microbial risk assessment of S. enterica related to mango, tomato, and raw chicken consumption in the central region of Mexico. PRACTICAL APPLICATION: Data on the prevalence and concentration of Salmonella enterica obtained in this study can be used to estimate the exposure assessment for the consumption of mango, tomato, and chicken in the central region of Mexico. In addition, the characteristics of the S. enterica isolates could be used to select representative strains for future studies to evaluate the intraspecies variability.

Modelling growth and histamine formation of Klebsiella aerogenes TI24 isolated from Indonesian pindang.

Indonesian salted-boiled fish (pindang) is a popular traditional food in Indonesia, which is made from Scombroid fish such as tuna and mackerel. As with other traditionally prepared fish products, pindang has important economic and social values, especially for those living in the coastal areas of Indonesia. However, pindang is a major cause of histamine fish poisoning (HFP) for consumers. Klebsiella aerogenes T124, a relatively high histamine-producing isolate from pindang, was used to describe lag time (λ), growth rate (µmax), maximum population density (Nmax), and histamine production in histidine broth and artificially contaminated Grey mackerel. Broth was adjusted to 1.5, 6, 10 and 20% w/v NaCl; mackerel was treated with 6% w/w NaCl, a level common to Indonesian industry practice, or not treated with additional NaCl. Samples were incubated at 10, 15, 20 and 30 °C. In broth, µmax and Nmax were significantly affected by temperature and NaCl, respectively, with λ influenced by both parameters. In control fish, µmax was significantly affected by temperature and NaCl, except at 10 and 15 °C; for 6% NaCl treatment, growth was only observed at 20 and 30 °C. Under similar incubation conditions for broth and fish, histamine formation was markedly affected by NaCl concentration. In broth, -5.1 to -6.6 log µg of histamine was produced per CFU, versus -4.6 to -6.6 log µg per CFU in fish. This study demonstrated that mackerel treated with 6% NaCl and stored at 10-15 °C prevents growth of K. aerogenes strain TI24 and formation of toxic levels of histamine.

Modelling viability of Listeria monocytogenes in paneer.

Paneer is a fresh, soft ready-to-eat cheese that is susceptible to Listeria monocytogenes contamination, exemplified by product recalls in Australia, Canada, and the USA. Previous research demonstrates that L. monocytogenes grows in paneer, however there are no paneer-specific predictive models that quantify the effect of environmental conditions on L. monocytogenes viability. This study measured the viability of a five-strain cocktail of L. monocytogenes in freshly prepared paneer incubated at 4-40 °C. Growth rates were fitted with the extended Ratkowsky square root model, with growth rates ranging from 0.014 to 0.352 log10 CFU/h. In comparison with published models, only the ComBase L. monocytogenes broth model acceptably predicted growth (Bf = 1.01, Af = 1.12) versus the developed model. The influence of paneer pH (5.0-6.0) and storage temperature (41-45 °C) on L. monocytogenes growth at the upper temperature growth boundary was described using a logistic model. These models provide quantitative tools to improve the safety of paneer processing conditions, shelf-life estimation, food safety management plans, and risk assessment.

Pathogen growth when implementing 'Time as a Public Health Control'.

Food regulatory authorities permit the use of Time as Public Health Control (TPHC) for handling foods that potentially support the growth of pathogenic bacteria. Considering the widespread use of TPHC in food service operations, few reports quantitatively describe potential pathogen growth when these protocols are implemented. A worst-case growth rate model was built from the highest growth rates predicted by ComBase broth-based models for six pathogens. A separate worst-case growth model was constructed from growth rates in ComBase database records. The maximum estimated pathogen growth in 4 h, assuming no lag phase, ranged from 0.006 log CFU at 5 °C to 6.16 log CFU at 44 °C, with 3.1 log CFU at 25 °C. In addition, pathogen growth when implementing TPHC could exceed the 1- and 3-log limits recommended for food challenge tests. The use of predictive models in development of TPHC criteria may provide more fail-safe strategies for managing microbial hazards in potentially hazardous food. This strategy could also reduce food waste and promote the use of temperature sensors in food supply chains.

Effect of glucose, pH and lactic acid on Carnobacterium maltaromaticum, Brochothrix thermosphacta and Serratia liquefaciens within a commercial heat-shrunk vacuum-package film.

Carnobacterium maltaromaticum, Brochothrix thermosphacta and Serratia liquefaciens are common spoilage organisms found within the microbiome of refrigerated vacuum-packaged (VP) beef. Extending and predicting VP beef shelf-life requires knowledge about how spoilage bacteria growth is influenced by environmental extrinsic and intrinsic factors. Multifactorial effects of pH, lactic acid (LA) and glucose on growth kinetics were quantified for C. maltaromaticum, B. thermosphacta and S. liquefaciens within a heat shrink-wrapped VP commercial film containing a simulated beef medium. LA, pH, and undissociated lactic acid (UDLA) significantly affected bacterial growth rate (p < 0.001), whereas 5.55 mM glucose produced a marginal effect. At 1.12 mM UDLA, growth rate and maximum population density decreased 20.9 and 3.5%, 56 and 7%, and 11 and 2% for C. maltaromaticum, B. thermosphacta, and S. liquefaciens, respectively.

qPCR quantification of Carnobacterium maltaromaticum, Brochothrix thermosphacta, and Serratia liquefaciens growth kinetics in mixed culture.

Quantifying growth kinetics of specific spoilage microorganisms in mixed culture is required to describe the evolution of food microbiomes. A qPCR method was developed to selectively amplify individual meat spoilage bacteria, Carnobacterium maltaromaticum, Brochothrix thermosphacta and Serratia liquefaciens, within a broth medium designed to simulate the composition of beef. An optimized method of DNA extraction was produced for standard curve construction. Method specificity was determined by individual single peaks in melt curves. Reaction efficiency for standard curves of C. maltaromaticum, B. thermosphacta and S. liquefaciens was high (R2 = 0.98-0.99), and linear quantification was achieved over a 5 log CFU/ml range. Coefficient of variation was calculated considering both threshold cycle (Ct) and bacterial concentration; the value did not exceed 14% for inter- or intra-runs for either method. Comparison of growth kinetic parameters derived from plate count and qPCR showed no significant variation (P > .05) for growth rate (GR) and maximum population density (MPD); lag phase duration (LPD) was not included in this comparison due to high innate variability. Log quantification of each isolate was validated in a mixed-culture experiment for all three species with qPCR and plate count differing less than 0.3 log CFU/ml (average 0.10 log CFU/ml, R2 = 0.98).

Salmonella enterica in Mexico 2000-2017: Epidemiology, Antimicrobial Resistance, and Prevalence in Food.

In Mexico, information of Salmonella enterica cases linked to food consumption is scarce. The objective of this article was to assess how S. enterica affect public health in Mexico. To conduct this study, data on the epidemiology of nontyphoidal S. enterica (NTS), Salmonella Typhi, and Salmonella Paratyphi A collected from 2000 to 2017 through the National Epidemiological Surveillance System of Mexico (Sistema Nacional de Vigilancia Epidemiológica de Mexico [SINAVE]) were used. Geographical distribution, season, age groups, and gender were variables considered to analyze S. enterica incidence. An estimation of cases caused by S. enterica in Mexico was calculated while considering data underestimation and the proportion of foodborne diseases. Information of the prevalence of the pathogen in food and the antimicrobial resistance of isolates from food and human cases were obtained from published studies. Outbreaks of S. enterica derived from imported Mexican products in the Unites States are discussed. In 2017, the numbers of reported cases of NTS (92,013) were two and seven times higher than the reported cases of Salmonella Typhi (45,280) and Salmonella Paratyphi A (12, 458). The NTS incidence was higher in lower socioeconomic Mexican regions. The gaps in the surveillance system make it impossible to establish a reliable tendency among age groups, geographical distribution, and gender. In 2017, the estimated frequency of NTS foodborne cases was 49 times higher than that reported in SINAVE, whereas for Salmonella Typhi and Salmonella Paratyphi A it was 23 times. Fresh meat showed the highest prevalence of S. enterica, and most of their isolates had multidrug resistance. Salmonella Typhimurium was the most common serotype isolated from human cases and food. Food safety agencies in Mexico need to prioritize efforts and resources to establish guidelines to ensure the absence of S. enterica in food.

Characterization of Bacterial Communities in Mexican Artisanal Raw Milk "Bola de Ocosingo" Cheese by High-Throughput Sequencing.

The dynamics of bacteria community of "Bola de Ocosingo" cheese, a Mexican artisanal raw milk cheese was investigated by high-throughput sequencing (454 pyrosequencing). Dairy samples (raw milk, curd, cheese at 50 and 110 days of ripening) were collected at dry (March-June) and rainy season (August-November) from three producers located in Chiapas, Mexico. In general, raw milk contained high bacterial diversity which was reduced throughout cheese manufacture. However, in two productions an important increase during cheese ripening was observed probably due to cross-contamination. Species such as Streptococcus thermophilus, Lactococcus lactis, Lactobacillus helveticus, L. delbrueckii and L. plantarum from which potential probiotic strains may be obtained, predominated during processing, varying its prevalence from one producer to another. Furthermore, low proportions of Escherichia coli/Shigella flexnerii were detected in almost all processes, however, could not be recovered by traditional methodology, indicating presence of non-cultivable cells. This work provides insights into bacteria communities of Bola de Ocosingo cheese for starter culture development, many of which are reported to provide health related benefits, and the usefulness of high-throughput sequencing to evidence cross-contamination during processing.

Integrating predictive models and sensors to manage food stability in supply chains.

Food products move through complex supply chains, which require effective logistics to ensure food safety and to maximize shelf-life. Predictive models offer an efficient means to monitor and manage the safety and quality of perishable foods, however models require environmental data to estimate changes in microbial growth and sensory attributes. Currently, several companies produce Time-Temperature Indicators that react at rates that closely approximate predictive models; these devices are simple and cost-effective for food companies. However, even greater outcomes could be realized using sensors that transfer data to predictive models in real-time. This report describes developments in predictive models designed for supply chain management, as well as advances in environmental sensors. Important innovation can be realized in both supply chain logistics and food safety management by integrating these technologies.

A Predictive Model for the Growth of Listeria monocytogenes in Commercial Blue Crab (Callinectes sapidus).

During the processing and handling of commercial blue crab (Callinectes sapidus), Listeria monocytogenes can potentially contaminate cooked meat and grow to hazardous levels. To manage this risk, predictive models are useful tools for designing and implementing preventive controls; however, no model specific for blue crab meat has been published or evaluated. In this study, a cocktail of L. monocytogenes strains was added to pasteurized blue crab meat, which was incubated at storage temperatures from 0 to 35°C. At selected time intervals, L. monocytogenes was enumerated by direct plating onto modified Oxford agar. A primary model was fitted to kinetic data to estimate the lag-phase duration (LPD) and growth rate (GR). Listeria monocytogenes replicated from 0 to 35°C, with GR ranging from 0.004 to 0.518 log CFU/h. Overall, the LPD decreased with increasing temperature, displaying a maximum value of 187 h at 0°C; however, this trend was not consistent. The LPD was not detected at 10°C, and it occurred inconsistently from trial to trial. A secondary GR model (R2 = 0.9892) for pasteurized crab meat was compared with the L. monocytogenes GR in fresh crab meat, demonstrating bias and accuracy factors of 0.98 and 1.36, respectively. The model estimates varied from other published data and models, especially at temperatures ≥5°C, supporting the need for a specific predictive tool for temperature deviations.

In vitro characteristics of an Atlantic salmon (Salmo salar L.) hind gut microbial community in relation to different dietary treatments.

In this study, microbial community dynamics were assessed within a simple in vitro model system in order to understand those changes influenced by diet. The abundance and diversity of bacteria were monitored within different treatment slurries inoculated with salmon faecal samples in order to mimic the effects of dietary variables. A total of five complete diets and two ingredients (plant meal) were tested. The total viable counts (TVCs) and sequencing data revealed that there was very clear separation between the complete diets and the plant meal treatments, suggesting a dynamic response by the allochthonous bacteria to the treatments. Automated ribosomal intergenic spacer analysis (ARISA) results showed that different diet formulations produced different patterns of fragments, with no separation between the complete diets. However, plant-based protein ingredients were clearly separated from the other treatments. 16S rRNA Illumina-based sequencing analysis showed that members of the genera Aliivibrio, Vibrio and Photobacterium became predominant for all complete diets treatments. The plant-based protein ingredient treatments only sustained weak growth of the genus Sphingomonas. In vitro based testing of diets could be a useful strategy to determine the potential impact of either complete feeds or ingredients on major fish gastrointestinal tract microbiome members.

Microbial and sensorial models for head-on and gutted (HOG) Atlantic Salmon (Salmo salar) stored from 0 to 15 °C.

Predictive models offer efficient means to manage the quality and safety of highly perishable seafood. Salmon is an increasingly popular seafood, and relies on well managed domestic and international supply chains to minimize growth of spoilage and pathogenic bacteria. While the literature describes predictive models for smoked and modified atmosphere packaged salmon, there are no reported models for spoilage bacteria and Listeria monocytogenes on head-on and gutted (HOG) aerobically-stored Atlantic salmon. Predictive models were developed for microbial and sensorial degradation of HOG Atlantic salmon stored at 0-15 °C until the end of shelf-life. Total Viable Count (TVC) and Pseudomonas spp. had similar growth rates at 0, 5 and 10 °C, but TVC rate was higher at 15 °C. L. monocytogenes growth rate at 0 °C was 0.004 log10 cfu/h, and showed a log-linear increase (R(2) = 0.99) to 0.079 log10 cfu/h at 15 °C. Sensory Quality Index (QI) scores were 2.4, 4.5, and 7.2 times greater at 5, 10 and 15 °C, respectively, compared to 0 °C. QI and TVC rates had a relatively strong relationship at 5 (R(2) = 0.87), 10 (R(2) = 0.80) and 15 °C (R(2) = 0.78), compared to 0 °C (R(2) = 0.50). These models are potential tools to manage the safety and quality of HOG Atlantic salmon in supply chains.

Atlantic Salmon (Salmo salar L.) Gastrointestinal Microbial Community Dynamics in Relation to Digesta Properties and Diet.

To better understand salmon GI tract microbial community dynamics in relation to diet, a feeding trial was performed utilising diets with different proportions of fish meal, protein, lipid and energy levels. Salmon gut dysfunction has been associated with the occurrence of casts, or an empty hind gut. A categorical scoring system describing expressed digesta consistency was evaluated in relation to GI tract community structure. Faster growing fish generally had lower faecal scores while the diet cohorts showed minor differences in faecal score though the overall lowest scores were observed with a low protein, low energy diet. The GI tract bacterial communities were highly dynamic over time with the low protein, low energy diet associated with the most divergent community structure. This included transiently increased abundance of anaerobic (Bacteroidia and Clostridia) during January and February, and facultatively anaerobic (lactic acid bacteria) taxa from February onwards. The digesta had enriched populations of these groups in relation to faecal cast samples. The majority of samples (60-86 %) across all diet cohorts were eventually dominated by the genus Aliivibrio. The results suggest that an interaction between time of sampling and diet is most strongly related to community structure. Digesta categorization revealed microbes involved with metabolism of diet components change progressively over time and could be a useful system to assess feeding responses.

High-throughput sequencing of microbial communities in Poro cheese, an artisanal Mexican cheese.

The bacterial diversity and structure of Poro cheese, an artisanal food, was analysed by high-throughput sequencing (454 pyrosequencing) in order to gain insight about changes in bacterial communities associated with the cheese-making process. Dairy samples consisting of milk, fermented whey, curd and ripened cheese (during 7 and 60 d) were collected from three manufacturers located in the state of Tabasco, México during dry (March-June) and rainy (August-November) seasons. Independently of producer and season, raw milk samples displayed the highest diversity in bacterial communities. In raw milk, genera found were Macrococcus, Staphylococcus, Enterococcus, Streptococcus, Lactobacillus and Enhydrobacter. Diversity in whey, curd and cheese was lower, principally containing Streptococcus and Lactobacillus; however, bacteria such as Staphylococcus, Acinetobacter, Chryseobacterium, Bacillus, Sediminibacter, Lactococcus and Enterococcus were occasionally present. After curdling step, the most dominant and abundant species were Streptococcus thermophilus and Lactobacillus delbrueckii.

Preliminary stochastic model for managing Vibrio parahaemolyticus and total viable bacterial counts in a Pacific oyster (Crassostrea gigas) supply chain.

Vibrio parahaemolyticus can accumulate and grow in oysters stored without refrigeration, representing a potential food safety risk. High temperatures during oyster storage can lead to an increase in total viable bacteria counts, decreasing product shelf life. Therefore, a predictive tool that allows the estimation of both V. parahaemolyticus populations and total viable bacteria counts in parallel is needed. A stochastic model was developed to quantitatively assess the populations of V. parahaemolyticus and total viable bacteria in Pacific oysters for six different supply chain scenarios. The stochastic model encompassed operations from oyster farms through consumers and was built using risk analysis software. Probabilistic distributions and predictions for the percentage of Pacific oysters containing V. parahaemolyticus and high levels of viable bacteria at the point of consumption were generated for each simulated scenario. This tool can provide valuable information about V. parahaemolyticus exposure and potential control measures and can help oyster companies and regulatory agencies evaluate the impact of product quality and safety during cold chain management. If coupled with suitable monitoring systems, such models could enable preemptive action to be taken to counteract unfavorable supply chain conditions.

Cultured C2C12 cell lines as a model for assessment of bacterial attachment to bovine primary muscle cells.

The mechanisms of bacterial attachment to meat tissues need to be understood to enhance meat safety interventions. However, little is known about attachment of foodborne pathogens to meat muscle cells. In this study, attachment of six Escherichia coli and two Salmonella strains to primary bovine muscle cells and a cultured muscle cell line, C2C12, was measured, including the effect of temperature. At 37°C, all but one strain (EC623) attached to C2C12 cells, whereas only five of eight strains (M23Sr, H10407, EC473, Sal1729a and Sal691) attached to primary cells. At 10 °C, two strains (H10407 and EC473) attached to C2C12 cells, compared to four strains (M23Sr, EC614, H10407 and Sal1729a) of primary cells. Comparing all strains at both temperatures, EC614 displayed the highest CFU per C2C12 cell (4.60±2.02CFU/muscle cell at 37 °C), whereas greater numbers of M23Sr attached per primary cell (51.88±39.43CFU/muscle cell at 37 °C). This study indicates that primary bovine muscle cells may provide a more relevant model system to study bacterial attachment to beef carcasses compared to cell lines such as C2C12.

Bacterial attachment to immobilized extracellular matrix proteins in vitro.

Meat surfaces are contaminated with bacteria during slaughter and processing. Understanding bacterial attachment properties to specific structures of meat could result in more targeted interventions to improve its safety and quality. However, the influence of temperatures relevant to abattoir environments on bacterial attachment to specific meat structures is not known. In this study, the effect of temperature and protein concentration on attachment of 10 Escherichia coli and seven Salmonella strains to extracellular matrix (ECM) proteins (collagen I, fibronectin, collagen IV and laminin) was measured using crystal violet stain and epifluorescence microscopy assays. By crystal violet assay, only five of 17 strains showed significant attachment to any ECM protein and only one strain attached to all proteins. Strains that attached at all tested temperatures (4, 25, 37°C) were E. coli M23Sr and M23 (collagen I); E. coli M23Sr (fibronectin); E. coli M23Sr, O157:H12 and M23, (collagen IV); and E. coli M23Sr, O157:H12, O78:K80:H1, O26:H11 and M23 (laminin). A higher proportion of strains attached to basement membrane proteins (laminin and collagen IV) than to interstitial proteins (collagen I and fibronectin). Highest attachment levels occurred at 4°C for collagen I and at 25°C for the other three proteins. Generally, the attachment levels of Salmonella strains to all ECM proteins were lower than for E. coli. No significant effect was found for concentration of collagen I, fibronectin and collagen IV, but was for higher laminin concentration. A strong positive correlation was found between results of both the crystal violet and epifluorescent methods (r≥0.905, p<0.05). This study demonstrated that attachment properties to ECM proteins displayed distinct variation among strains, that temperature highly influenced attachment and that protein concentration had a minor effect.

Prevalence, characterization and sources of Listeria monocytogenes in blue crab (Callinectus sapidus) meat and blue crab processing plants.

Seven blue crab processing plants were sampled to determine the prevalence and sources of Listeria spp. and Listeria monocytogenes for two years (2006-2007). A total of 488 raw crabs, 624 cooked crab meat (crab meat) and 624 environmental samples were tested by standard methods. Presumptive Listeria spp. were isolated from 19.5% of raw crabs, 10.8% of crab meat, and 69.5% of environmental samples. L. monocytogenes was isolated from 4.5% of raw crabs, 0.2% of crab meat, and 2.1% of environmental samples. Ninety-seven percent of the isolates were resistant to at least one of the ten antibiotics tested. Eight different serotypes were found among 76 L. monocytogenes isolates tested with the most common being 4b, 1/2b and 1/2a. Automated EcoRI ribotyping differentiated 11 ribotypes among the 106 L. monocytogenes isolates. Based on ribotyping analysis, the distribution of the ribotypes in each processing plant had a unique contamination pattern. A total of 92 ApaI and 88 AscI pulsotypes among the 106 L. monocytogenes isolates were found and distinct pulsotypes were observed in raw crab, crab meat and environmental samples. Ribotypes and serotypes recovered from crab processing plants included subtypes that have been associated with listeriosis cases in other food outbreaks. Our findings suggest that molecular methods may provide critical information about sources of L. monocytogenes in crab processing plants and will augment efforts to improve food safety control strategies such as targeting specific sources of contamination and use of aggressive detergents prior to sanitizing.