Introduction
Meat products are important contributors to food waste, with an estimated median of 9% and 11% of meat and meat products produced in the United States each year discarded at retail or household levels, respectively (FAO, 2025). Food waste occurring in the retail markets can occur by the use of conservative sell-by or use-by dates placed on the product by a manufacturer, used by consumers as a definitive marker of food safety (Gray, 2019; Xie and Bagchi, 2024). It is important to understand the factors impacting shelf life to ensure food safety and consumer satisfaction, as the shelf life is usually longer than what is indicated on the package (Palanisamy et al., 2024). These dates are just estimates of spoilage and relate more to quality deterioration than food safety marker (Jafarzadeh et al., 2024). Although many factors influence the shelf life of meat products, microbial spoilage is the primary determinant, and this is driven by the growth and metabolic activity of specific microbial communities present during storage (Chen et al., 2024). When the meat begins to spoil, metabolites are formed, which leads to changes in the characteristics of the meat including those that consumers use to identify safety and quality like color and odor (Odeyemi et al., 2020). The bacteria that are populating the surface of the meat product break down the tissue and produce the chemical compounds that make up the off odors and flavors of spoiled meat, in addition to color changes, sliming, and gas production. Understanding the initial microbial community structure can provide insights into how the product microbially spoils and what taxa are causing it. By linking the initial microbial spoilage indicators, such as sensory changes, it may be possible to develop ways to better predict the shelf life (Altaf and Ksouri, 2025). This approach allows for better management of the meat supply chain and less food waste at the consumer and retail levels.
Diversity of microbial communities change over the shelf life of ground beef, and the diversity significantly decreases throughout shelf life; however, the role microbial communities play in causing spoilage is crucial (Gram et al., 2002; Odeyemi et al., 2020). Specific spoilage organisms such as Pseudomonas spp. and lactic acid bacteria (LAB) play a role in degrading meat quality and causing off odors and discoloration associated with spoilage (Pellissery et al., 2020). More recent studies have demonstrated that it is not just the organisms operating individually that affect the spoilage patterns but the microbial community and microbial interactions influencing it (Fidan et al., 2022). Moreover, it is likely that the microbial community is changing over time during spoilage in a predictable succession pattern, as has been demonstrated in other decomposition environments (Burcham et al., 2024). Identifying these spoilage communities can aid in estimating the length of time available before the product is microbially spoiled. Being able to identify these spoilage communities and model the changes in community over time may help generate better predictions of microbial spoilage outcomes (Mamat et al., 2024).
Ultimately, a more thorough understanding of microbial spoilage of meat will provide a scientific basis for dynamic shelf-life labeling and reducing the environmental impact of food waste. Therefore, the objective of this study was to identify changes in the microbial community composition over the shelf life of beef products and associate this with key changes in meat quality and spoilage markers. This work shows that there is a predictable pattern of microbial succession in the spoilage community, which could be harnessed to better anticipate spoilage rates for decisions regarding product labeling.
Materials and Methods
Experimental design
This study was conducted as a longitudinal observational experiment designed to determine the relationship between microbial community structure and beef spoilage. Ground beef production lots were used as the experimental subject for this experiment, as they facilitate repeated destructive sampling of a consistent product over the beef shelf life. Ground beef samples from 3 grinding lots (n = 3) were partitioned into subsamples for daily destructive sample collection. One subsample from each sample (lot) was collected daily for 14 d, including day 0 as the day of placement. These samples were used to analyze color, temperature, weight, nutrient composition, and aerobic plate counts (APC). After initial sampling, the remaining product was frozen at −18°C (FKFH21F7HW Frigidaire, Charlotte, NC, USA) for DNA extraction and thiobarbituric acid-reactive substances (TBARS) analysis to be conducted later in the experimental period.
Raw materials
Boxed beef shoulder clods (US Department of Agriculture institutional meat purchasing specification #114) were purchased from a commercial beef processing facility in the high plains region of the United States and commercially transported to the Auburn University Lambert-Powell Meat Laboratory under refrigerated conditions (1.5°C). The beef was ground using a 9.525-mm plate, followed by another grind with a 3.175-mm plate using a commercial meat grinder (Model AFMG-48, Biro Manufacturing Company, Marblehead, OH, USA). The grinder was disassembled, cleaned, and sanitized between grinding lots to simulate a different operation to produce 3 lots of ground beef. From each lot, subsamples were weighed out to 225 g, tray overwrapped, and randomly assigned sampling pull dates (N = 45). The samples were then placed into simulated retail storage in an LED-lit (2297 lux; ILT10C, International Light Technologies, Peabody, MA, USA), 3-tier display open-curtain retail case (Model TOM-60DX-BN, Turbo Air Inc., Long Beach, CA, USA). To best simulate retail storage, the curtain was closed at night and open during the day. To prevent bias by cooler position, samples were randomly shuffled daily throughout the experiment as recommended by the American Meat Science Association (AMSA) Guidelines for Meat Color Measurement (King et al., 2023). Case temperature was monitored via a built-in thermometer throughout the day to ensure proper temperature control (3.0°C).
Instrumental color
Instrumental color analysis was measured with the HunterLab MiniScan EZ colorimeter, model 45/0 LAV (Hunter Associates Laboratory Inc., Reston, WV, USA), following the AMSA Meat Color Measurement Guidelines (King et al., 2023). The colorimeter was standardized using a black and white tile cover prior to each use and analyzed through the overwrap packaging. Instrumental color values were determined from the average of 3 readings per ground beef sample. Additionally, photographs were taken of each sample to monitor visible color changes during storage.
Compositional analysis
A meat compositional analysis was completed during sampling to determine protein moisture and fat content in each sample. The analysis was conducted using a near infrared with AOAC (2007.4) approved spectrophotometer (FoodScan™, FOSS Analytical A/S, Hilleroed, Denmark), and data processing was determined using ISIscan™ software to analyze the nutrient content in the meat (Martin et al., 2013).
Aerobic plate counts
APC were completed each day of storage to determine the rate of microbial growth. A 1.0-g portion of meat, representative of from the top, bottom, and inside of the sample, was placed in a sterile whirl-pak filter bag (Whirl-Pak, Pleasant Prairie, WI, USA). Sterile phosphate buffer (Criterion Dehydrated Culture Media, Hardy Diagnostics, Cat no. C9241, Santa Maria, CA, USA) (10 mL) was added to the meat and hand massaged for 60 s. Serial dilutions were done to measure aerobic microorganisms present in the samples. Dilutions were spread on tryptic soy agar media and incubated at 36°C for 48 h before counting (Criterion Dehydrated Culture Media, Hardy Diagnostics, Cat no. C7141, Santa Maria, CA, USA). Plate counts were log transformed and reported as log CFU/g. The plate counts demonstrated when the samples reached spoilage, indicated by 7-log CFU/g (Alexa et al., 2024).
Lipid oxidation
The development of oxidative rancidity and lipid oxidation in the samples was evaluated throughout the sampling period using 2-TBARS. During each day of sampling, 2 g from each ground beef sample were frozen (−18°C) for later analysis. Samples were frozen and stored for less than 7 d postexperiment. This method provides the most accurate TBARS analysis to ensure methodological consistency (Aguilar Diaz De Leon and Borges, 2020). At the completion of all sampling days, the samples that were frozen after initial analyses were thawed and homogenized (Model PT 10-35, Brinkman Inst., Westbury, NY, USA) into a uniform sample in duplicates and mixed with 8 mL of chilled phosphate buffer (pH 7.0), containing 0.1% ethylenediamine-tetraacetic acid, 0.1% n-propyl gallate, and 2 mL of trichloroacetic acid. The homogenized samples were filtered through a Whatman No. 1 filter paper into glass tubes, and duplicate 2-mL aliquots of clear filtrate were transferred into 10-mL test tubes. Filtrate was mixed with 2 mL of 0.02 M 2-thiobarbituric acid reagent and placed into a hot water bath (100°C) for 20 min. Following the hot water bath, the tubes were transferred to an ice bath for 15 min. The absorbance of each sample was measured at 533 nm with a spectrophotometer (Model UV-1600PC, VWR International LLC, Randor, PA, USA). The values were measured by milligrams of malonaldehyde per kilogram of fresh meat and multiplied using the value of 12.21 according to the standard curve to obtain the TBARS value (Buege and Aust, 1978).
Microbiome analysis methods
After all samples were collected, the 2.0-g frozen ground beef samples separated for microbiome analysis were thawed for DNA extraction. Prior to extraction, a 2-step centrifugation pre-extraction treatment was completed on the samples prior to DNA isolation to limit the host (bovine) DNA associated with the analysis, as recommended by previous literature (Hultman et al., 2015). Samples were diluted in 10 mL of peptone buffer and hand massaged for 30 s. Supernatant from the homogenization was then placed in a 15-mL conical tube and centrifuged at 200 rcf for 3 min (Eppendorf Centrifuge 5810 R). Following the first centrifugation, 10 mL of the supernatant were removed and centrifuged again at 10,000 rcf for 3 min (Eppendorf Centrifuge 5425) (Hultman et al., 2015). DNA was then extracted from the final pellet using Zymo Research Quick-DNA™ Miniprep Kit (Lot No. 237579, Irvine, CA, USA), following manufacturer instructions with the optional steps for low biomass. DNA quantification was done with an Invitrogen Qubit™ Flex Fluorometer.
Extracted DNA from all samples was frozen and shipped on dry ice to Novogene Corporation Inc. (Sacramento, CA, USA) for amplicon sequencing of the V3-V4 region of the ribosomal RNA (rRNA) gene at their China laboratory location (Guangzhou City, Guangdong Province, China). Sequencing was conducted on all samples in 1 lane on the Illumina NovaSeq platform with paired-end 250-bp chemistry (Illumina Inc., San Diego, CA, USA). Microbial 16S rRNA gene amplicons were analyzed using QIIME2 version 2024.2 (Bolyen et al., 2019). The sequences were denoised, and paired reads were joined using DADA2 (Callahan et al., 2016). Taxonomic classification was done using the SILVA 138-99 pretrained classifier with the QIIME feature-classifier plugin (Hung et al., 2022; Bokulich et al., 2018; Quast et al., 2013). Nonmicrobial DNA (chloroplast, mitochondria) was filtered from the dataset. Differential abundance analysis was conducted using the analysis of composition of microbiomes (ANCOM) with bias corrections, with Holm–Bonferroni multiple-test correction to suggest taxa with significant changes (Lin and Peddada, 2020). Alpha diversities (observed features, Faith’s phylogenetic diversity, and Shannon’s diversity) were statistically compared using a Kruskal–Wallis test, with Benjamini–Hochberg multiple-test correction (Faith, 1992). Beta diversities were analyzed using the Bray–Curtis dissimilarity metric, unweighted UniFrac, and principal coordinates analysis (PCoA) analysis was analyzed through QIIME2 Emperor (Lozupone et al., 2011). Furthermore, changes seen throughout the shelf life were calculated using the QIIME2 longitudinal plugin (Bokulich et al., 2018) with the volatility and linear mixed-effects tests (Xia and Sun, 2023; Xia et al., 2025), and the statistical assessments were made using the permutational multivariant analysis of variance test (Lozupone and Knight, 2005; Anderson, 2017).
The predictability of microbial succession during ground beef spoilage was assessed using a machine learning model constructed using the random forest algorithm (Zhang and Lu, 2012; Metcalf et al., 2015). The random forest algorithm is appropriate for this situation to manage the large number of predictive features while reducing, as much as possible, the likely overfitting issue due to a limited samples size. The microbiome relative abundance data were used as the predictor variable, with spoilage metric as the response variable. Models were trained to predict the following: (1) the day of retail display (day of spoilage), (2) APC levels, and (3) TBARS value. These were constructed using a K-fold cross-validation with hyperparameter tuning, as is appropriate for a situation with small sample sizes to reduce overfitting (Papoutsoglou et al., 2023). Grouping was applied such that all microbial data within a sample (production lot) of ground beef were placed in a validation fold together, and a random state parameter was applied to support feature stability across the folds. The accuracy of the models was assessed during cross-validation using the mean absolute error (MAE), defined as the deviation of the predicted from observed values represented in the same unit as the original data ([1] days; [2] log CFU/g; [3] mg of malonaldehyde/kg). Recursive feature elimination was used to identify the number and name of features (microbial amplicon sequence variants [ASV]) that contributed to the most accurate final model. For each model, the taxonomic names associated with the important ASV were assigned for comparison with other analyses. All modeling was conducted using the python machine learning package scikit-learn version 19.0 (Pedregosa et al., 2011). Results were visualized using ggplot2 in R (Wickham, 2016).
Statistical analysis
Statistics for nonmicrobiome analyses were analyzed using R Studio version 4.2.2 using the packages emmeans, tidyverse, and ggplot2 (R Core Team, 2022; Wickham, 2016; Wickham et al., 2019; Lenth et al., 2025). Means were compared using linear mixed-effects models with day as a fixed effect and sample as a random effect. Significance was identified at a P value less than .05. Statistics for microbiome analysis were analyzed using QIIME 2 and visualized using the ggplot2 package in R Studio version 4.2.2 (Bolyen et al., 2019; R Core Team, 2022; Wickham, 2016).
Results
Ground beef quality
This experiment monitored the quality and microbial changes in ground beef products over the shelf life to understand the microbial succession patterns. Overall, quality data demonstrated progressive visual color changes and spoilage patterns throughout the experimental period (Supplemental Figure 1). As expected, there was little variation in the meat compositional analysis over the experimental period. There was significance among protein (P < .01; CI = 0.19–0.32) and fat (P < .01; CI = 0.17–0.27) by day, but no changes in moisture (P = .622) possibly due to the aerobic packaging and oxygen availability (Supplemental Table 2). The compositional analysis shows a statistical difference over time, but this shift is not biologically significant and does not impact product quality. Instrumental color, APC, and TBARS were monitored as a metric for product spoilage. There were no differences (P > .05) for full analysis of instrumental color, APC, or TBARS among samples of ground beef per each day of simulated storage. There was a decrease (P < .01; CI = 1.44–2.43) in the surface lightness (L*), redness (a*), and yellowness (b*) throughout storage until day 9, at which point the color became stabilized (Figure 1A, Supplemental Table 1). APC increased (P < .01; CI = 0.38–0.47) over time, crossing the threshold for the definition of spoilage (7 log CFU/g) on day 6 for all samples (Figure 1B). Similarly, TBARS, quantified by milligrams per kilograms of malonaldehyde, showed an increasing trend over time. The TBARS increased (P < .01; CI = 0.61–0.97) from day 0 to day 8, then decreased (P < .01) until day 11, before increasing until the end of the experimental period. TBARS crossed the threshold for spoilage defined by 2.0 mg/kg of malonaldehyde by oxidative rancidity on day 6 (Figure 1B) (Hernández-García et al., 2022).
Ground beef quality deterioration through the 14-d experimental period, assessed by linear mixed-effects models with day as a fixed effect and sample as a random effect (n = 3/d). A) Instrumental color analysis showing changes in L* (lightness; P < .01), a* (redness; P = .007), and b* (yellowness; P < .01) values, all measures decreased over time. Ribbons indicate standard error. B) APC showing changes in plate counts throughout the experimental period (P < .01), and lipid oxidation (TBARS) analysis showing changes in milligrams per kilograms malonaldehyde (P < .01), all following expected spoilage progression. APC, aerobic plate counts; TBARS, thiobarbituric acid-reactive substances.
Results from the analysis of composition of microbes with bias correction and a Holm–Bonferroni multiple-test correction.
| Feature ID (Family) | Log Fold Change | W Score |
|---|---|---|
| Morganella | 5.38 | 13.10 |
| Vagococcus | 5.29 | 8.48 |
| Enterobacterales | 4.91 | 9.97 |
| Morganellaceae | 4.90 | 8.58 |
| Yersinia | 5.53 | 7.27 |
| Enterobacteriaceae | 4.36 | 9.37 |
| Yersiniaceae | 3.97 | 11.85 |
| Yersiniaceae | 3.81 | 11.85 |
| Providencia | 3.68 | 6.02 |
| Providencia | 3.60 | 5.86 |
| Moellerella | 3.46 | 6.82 |
| Enterobacterales | 3.31 | 6.66 |
| Serratia | 3.29 | 5.48 |
| Oxalobacteraceae | 2.45 | 5.11 |
| Moellerella | 2.21 | 4.41 |
The analysis determines differentially abundant organisms between the pre- and postspoilage states (before and after day 6 of 14 d) as defined by aerobic plate counts and lipid oxidation levels (N = 45 subsamples; P < .001; Q < 0.001).
Microbiome analysis
The microbiome of ground beef throughout the shelf life was evaluated using 16S rRNA microbial gene sequencing. Sequencing a total of 45 samples resulted in a total of 8,951,852 demultiplexed sequences. The reads were denoised into 656 ASV, with 641 ASV remaining after filtering for nonmicrobial taxa (mitochondria and chloroplast). Samples were rarefied at 105,506 sequences for diversity analysis, using 35 out of 45 rinsate samples.
There were no differences among the samples for richness (mean = 39.9; P = .097), Shannon’s diversity (mean = 3.44; P = .115), and Faith’s diversity (mean = 13.1; P = .113) (Figure 2A). However, β diversity, as measured by unweighted UniFrac distances, increased over time (P = .009). Furthermore, there was a clear difference in β diversity between unspoiled and spoiled samples, as defined by APC (P < .05) (Figure 2B). When clustered using PCoA analysis, UniFrac dissimilarity shows clustering based on the spoilage state by each experimental day (Supplemental Figure 2).
Microbial community analysis throughout the 14-d experimental period. A) Alpha diversity over days of spoilage. Samples were compared by Kruskal–Wallis test with Benjamini–Hochberg multiple-test correction (n = 3/d). There was no difference (P > .05) for any diversity metric. B) Unweighted UniFrac PCoA showing clustering based on spoilage state (shape) and by day (color). C) Taxonomic representation of bacterial families present throughout the experimental period, showing a higher number of microbial communities on the day of grinding and changes throughout the experimental period. PCoA, principal coordinates analysis.
Taxonomy
These shifts in β diversity are also shown by the taxonomy. Ground beef has an initial microbiome that changes throughout the shelf life. As the shelf life progressed, the most dominant increase was in Pseudomonadaceae, which is a known leading spoilage organism (Taormina, 2021). There are also higher populations of Carnobacteriaceae, Listeriaceae, Streptococcaceae, Lactobacillaceae, and Yersiniaceae throughout the shelf life. Although many organisms are present on the day of grinding, there is a decrease in the number of bacterial families present throughout the experiment, as the organisms outcompete each other in this time, fighting for limited resources (Figure 2C). Additionally, when analyzing the differential abundance through ANCOM, there were significant changes among the taxa that are significantly enriched in the spoiled condition compared to the nonspoiled reference samples (P < .01). Results of ground beef samples demonstrate the top-enriched taxa displayed the greatest log fold change during storage from day 0 to 14 (Table 1).
Machine learning
Changes in the microbial community composition were predictable. Machine learning models were constructed and validated using K-fold cross-validation with hyperparameter tuning to determine the association of the microbiome with the changes in ground beef quality throughout shelf life. Random forest regression models using the microbial community with 642 microbial features as the predictor and the quality indicator and 3 samples per day measured for quality as the response variable were successfully constructed and evaluated. The MAE for the models were 1.09, 2.98, and 2.48 for predicting APC, TBARS, and days of spoilage, respectively. The most informative features (ASV) for each model were assigned taxonomic identifiers (Figure 3). The APC most optimal model contained 15 features, which included Methyloceanibacter, uncultured Micropepsaceae family, Flavobacteriaceae, Prevotella, Puniceicoccaceae, Enterococcaceae, Yersinia, Pseudomonadales clade SAR86, Acidobacteriaceae subgroup 1, Puniceicoccaceae, Latilactobacillus, Carnobacterium, Akkermansia, Varibaculum, and Longimicrobiaceae. The TBARS most optimal model contained 6 features, which included Methyloceanibacter, uncultured Micropepsaceae, Flavobacteriaceae, Prevotella, Puniceicoccaceae, and Enterococcaceae. The day model most optimal model contained 9 features, which included Methyloceanibacter, uncultured Micropepsaceae, Flavobacteriaceae, Prevotella, Puniceicoccaceae, Enterococcaceae, Yersinia, Pseudomonadales SAR86 clade, and Acidobacteriaceae subgroup 1.
Important features identified by random forest regression analysis. Random forest models were trained using K-fold cross-validation with hyperparameter tuning, and the most accurate models were selected after recursive feature selection. Models were trained with microbial community (amplicon sequence variants) as the predictor variable and a measure of microbial spoilage (APC, experimental day [day of spoilage], lipid oxidation [TBARS]) as the response variable. The number of features identified in the most accurate model were selected, and the features with the greatest contribution to each model are identified here. APC, aerobic plate counts; TBARS, thiobarbituric acid-reactive substances.
Discussion
The results of this study align with the expected outcomes of a typical observational spoilage experiment, making it an ideal model for the research question. While the small sample size (n = 3/d) limits the interpretation of results, these data demonstrate a predictable microbial succession pattern during spoilage that can be further expanded upon in future work. This study shows how the microbial communities associated with meat product spoilage play a large role in the rate of spoilage and how the communities change over the shelf life.
There is evidence, according to APC, TBARS, and instrumental color testing, which supports spoilage occurring on day 6 of the experimental period. The instrumental color data provided in this study support the results of microbial activity throughout the shelf life. The consistent decrease of L*, a*, and b* values until microbial spoilage demonstrates how the microbes play a role in the degradation of myoglobin (Hu et al., 2024). The color degradation shown in the current study provides an accurate depiction as to how ground beef would change when packed aerobically in a retail environment. The high-permeability tray overwrap packaging did influence the shift in color and other quality characteristics, as the availability of oxygen is an important factor in the myoglobin oxidation process. As the packaging was aerobic, APC were the appropriate method to identify microbial growth patterns and to establish the date of microbiological spoilage, though microbiome analysis presented below indicates growth of anaerobic organisms as well (Chen et al., 2014).
Lipid oxidation data were analyzed in this study to determine oxidative rancidity throughout shelf life and the relations to microbial community activity. Spoilage from this mechanism aligned closely with the shifts in myoglobin state, reflecting meat color and the microbial spoilage APC levels observed on day 6 of the experiment. Previous works have shown lipid oxidation and observed increasing TBARS values throughout storage time in frames similar to those presented here (Li et al., 2025). Li et al. (2025) specifically demonstrated an abundance of organisms, such as Pseudomonas spp., associated with the increase in TBARS throughout the shelf life, which is a driver for spoilage and links the presence of microbes to rancidity. In the current study, a link between the changing TBARS values and shifts in the taxonomy was also identified. The TBARS values did not increase consistently, with a sudden decrease in values between days 9 and 11. Interestingly, this was accompanied by a shift in the relative abundance of Pseudomonas and LAB. Over the same period as the presence of reactive substances decreased, the relative abundance of the LAB increased similarly. While the presence of Pseudomonas spp. has been previously linked to increased lipid oxidation and the formation of secondary oxidation products, more research is needed to elucidate the exact nature of this relationship (Fang et al., 2022). Interestingly, LAB has been shown to produce some antioxidant effects in processed meat products; therefore, the outgrowth of these organisms may contribute to the shift in TBARS, though it is currently unclear why this would occur at this point in spoilage (Xia et al., 2023).
Microbiome analysis was conducted in this study with the aim of identifying patterns in the microbial community over the shelf life of meat products. In this case, results suggested a predictable pattern that was identifiable across samples of ground beef. The initial microbiome was similar across ground beef samples, so it was difficult to discern the impact of this initial community on spoilage conditions. However, previous work demonstrates that the early microbiome does shape the spoilage patterns, and future work can further evaluate this over the longitudinal period (Hwang et al., 2020). Most previous work identifying major aerobic spoilage organisms identify Pseudomonas spp. as the major drivers (Wickramasinghe et al., 2019; Snyder et al., 2024). Hwang et al. (2020) demonstrated that a diverse initial microbial community will quickly become dominated by Pseudomonas as spoilage progresses. Previous work has demonstrated the species of Pseudomonas drives different spoilage patterns (Stanborough et al., 2018); however, in the current study, limitations of the 16S rRNA methodology do not allow for species-level resolution to confirm this. The number of different ASV assigned to the Pseudomonadaceae family, though, does imply some species diversity that can be further investigated in future work: the taxonomic patterns identified in the current study.
The major change in microbial community over the shelf life of ground beef was associated with a shift in the relative abundance of microbial taxa rather than a shift in the overall diversity. On day 0, there were many microbial species represented, many at very low abundance. Immediately after the onset of spoilage, the taxonomy shifted to show a reduced number of species, which were mostly those previously associated with spoilage. This decrease in taxonomic diversity did not align with the α diversity results, possibly due to many ASV assigning to similar taxa or a statistical artifact of low sample sizes. The shift in taxonomy aligns with previous work as well. Xiao et al. (2025) also showed an abundance of Pseudomonas throughout the experimental period, indicating that these organisms are strongly associated with the spoilage process. Additional research completed by Asmus et al. (2025) included similar microbiome analysis methods; however, this study was done in pork and included a mock community. The study, which looked at microbial profiling of fresh pork cuts, found the microbial profile to be significantly different and that production facilities can have an influence on the fresh pork microbiome. Due to the addition of the mock community, a lower abundance of the spoilage organisms was present throughout shelf life as they were outcompeted. The changes observed in microbial communities as spoilage progressed aligned with results presented here, though it was a different group of organisms influenced by the initial inoculation. To underscore the relationship between microbial profile and spoilage progression, ANCOM analysis was used to determine the differentially abundant organisms between spoiled and unspoiled products. Spoiled products contained a greater relative abundance of ASV aligned to Morganella, Vagococcus, Enterobacteriales, and Yersinia. Previous research has identified these organisms as contributors to the spoilage of meat products (Al-Mazrouei et al., 2024; Xu et al., 2025).
The outcomes of the taxonomic, ANCOM, and machine learning analyses provide evidence for microbial succession. There is clear overlap with the taxonomic changes and the microbial communities associated with spoilage. The taxonomic results of this study show trends in Carnobacteriaceae and Pseudomonadaceae throughout the experimental period. Likewise, results from the random forest regression model show that Carnobacterium, among other organisms, is important for predicting day of spoilage based on APC data. The machine learning analysis showed that the communities that contribute to ground beef spoilage are predictable. A similar study demonstrated the use of machine learning in detection of Escherichia coli O157:H7 in ground beef and found that the random forest model was able to predict the growth performance based on different temperatures (Al et al., 2024). The researchers were able to use regression models to determine the growth rate of E. coli O157:H7 and its predictability, providing valuable insight into the food safety industry. Similarly, the current study was able to monitor the changes in microbial communities and prove the concept that communities change as spoilage progresses. The use of artificial intelligence in the food industry is allowing an increase in understanding, which can lead to a safer and more evolved industry (Jayan et al., 2025). Research has also shown that using this technology to understand microbial shelf life and predictability offers benefits to sustainability and the food supply chain (Tarlak, 2023). In this study, the sample size used for model training was very low (n = 3), which limits the generalizability of the results and diminishes the utility of the model beyond the current dataset. The random forest algorithm, supported by K-fold cross-validation and hyperparameter tuning, was selected as a method to reduce potential overfitting, but, without a deeper dataset and external validation, it is likely the presented models are overfit (Papoutsoglou et al., 2023). However, the objective of this manuscript is to demonstrate the potential for a machine learning tool to evaluate shelf life based on the microbiome, which is evident here. The research done here shows there is a predictable pattern to follow to understand these microbial changes throughout spoilage, an outcome that can be expanded in future work.
Limitations and future work
This research study had many aspects that allowed for a new look into microbial meat spoilage, and we were able to see the different microbial communities changing throughout the shelf life of ground beef. The greatest limitation of this study was in the small sample number, which allowed preliminary findings in the area but not generalizable conclusions. It is difficult to define an optimal sample size for microbiome studies, especially when machine learning is employed; however, now that this study has demonstrated relatively large effects sizes associated with spoilage shelf life, these can be calculated with power analysis for future study. The sample size generally used to define the threshold for a “small study” in machine learning is 100 samples per treatment, which may provide practical challenges in conducting a study of this type. The use of cross-validation methods can help ameliorate this challenge, and the work could be done over several time periods to allow more samples even with limited physical space. Despite this limitation, the results presented here will serve as a basis for future work, investigating trends in microbial succession and their relationship with quality outcomes. Beyond increasing the sample size, more diverse sources for the meat samples that provide more variation in the initial microbiome may have allowed for more detailed results and models with less potential overfitting. Additional work may also employ other omics methods to investigate the fundamental trends underlying the observed patterns presented here.
Conclusion
The overall goal of this study was to observe the microbial spoilage patterns associated with ground beef to better predict microbial spoilage. The results conclude that through the use of quality results, ground beef achieves microbial spoilage by day 6 of refrigerated storage. The microbiome analysis and machine learning technology results of this experiment allow for a better understanding of ground beef spoilage and can help reduce food waste at the retail and consumer levels. Findings from this study align with previous results throughout the literature, highlighting changes in microbial communities throughout spoilage; however, further analysis is required to answer the question relating to predicting ground beef spoilage based on the initial microbial community profile.
Conflict of Interest
The authors of this manuscript declare no conflict of interest.
Acknowledgments
This research project was funded by the Alabama Beef Checkoff. The authors express gratitude for assistance with research to the support staff of the Lambert-Powell Meat Laboratory and graduate students Barrett Maloney, Savannah Douglas, and Gabriela Bernardez-Morales, along with the assistance with resources and writing feedback from Dr. Jason Sawyer.
Author Contribution
Isabella G. Gafanha: data collection, data analysis, investigation, writing—original draft, reviewing, and editing; Barney Wilborn: data collection, and writing—reviewing and editing; Dianna Bourassa: writing—reviewing and editing; Amit Morey: writing—reviewing and editing; and Aeriel D. Belk: conceptualization, funding acquisition, data analysis, investigation, principal investigator, writing—reviewing and editing.
Data Availability
Microbiome data generated during this experiment are available in the European Nucleotide Archive database as study PRJEB96000 (ERP178751) and on QIITA as study 16076. Coding scripts and python notebooks containing model construction methods are available on GitHub at https://github.com/aeriel-belk/shelf-life-microbiome.
Literature Cited
Aguilar Diaz De Leon, J., and C. R. Borges. 2020. Evaluation of oxidative stress in biological samples using the thiobarbituric acid reactive substances assay. J. Visualized Exp. 159:e61122. doi: https://doi.org/10.3791/61122
Al, S., F. U. Ciloglu, A. Akcay, and A. Koluman. 2024. Machine learning models for prediction of Escherichia coli O157:H7 growth in raw ground beef at different storage temperatures. Meat Sci. 210:109421. doi: https://doi.org/10.1016/j.meatsci.2023.109421
Alexa, E. A., A. Papadochristopoulos, T. O’Brien, and C. M. Burgess. 2024. Microbial contamination of food. In: A. K. Jaiswal and S. Shankar, editors, Food packaging and preservation: antimicrobial materials and technologies. Academic Press. p. 3–19. doi: https://doi.org/10.1016/B978-0-323-90044-7.00001-X
Al-Mazrouei, M. A., Z. S. Al-Kharousi, J. M. Al-Kharousi, and H. M. Al-Barashdi. 2024. Microbiological evaluation of local and imported raw beef meat at retail sites in oman with emphasis on spoilage and pathogenic psychrotrophic bacteria. Microorganisms. 12:2545. doi: https://doi.org/10.3390/microorganisms12122545
Altaf, Q. S., and R. Ksouri. 2025. Application of machine learning in the food industry. In: I. Zahoor, S. A. Wani, and T. A. Ganaie, editors, Artificial intelligence in the food industry: enhancing quality and safety. CRC Press, Boca Raton, FL.
Anderson, M. J. 2017. Permutational multivariate analysis of variance (PERMANOVA). In: N. Balakrishnan, T. Colton, B. Everitt, W. Piegorsch, F. Ruggeri, and J. Teugels, Wiley StatsRef: statistics reference online. John Wiley & Sons, Ltd. doi: https://doi.org/10.1002/9781118445112.stat07841
Asmus, A. E., T. N. Gaire, K. J. Schweisthal, S. M. Staben, and N. R. Noyes. 2025. Microbiome characterization of two fresh pork cuts during production in a pork fabrication facility. Microbiol. Spectrum. 13:e02209–24. doi: https://doi.org/10.1128/spectrum.02209-24
Bokulich, N. A., B. D. Kaehler, J. R. Rideout, M. Dillon, E. Boylen, R. Knight, G. A. Huttley, and J. G. Caporaso. 2018. Optimizing taxonomic classification of marker-gene amplicon sequences with QIIME 2’s Q2-feature-classifier plugin. Microbiome. 6:90. doi: https://doi.org/10.1186/s40168-018-0470-z
Bolyen, E., J. R. Rideout, M. R. Dillon, N. A. Bokulich, C. C. Abnet, G. A. Al-Ghalith, H. Alexander, E. J. Alm, M. Arumugam, F. Asnicar, Y. Bai, J. E. Bisanz, K. Bittinger, A. Brejnrod, C. J. Brislawn, C. T. Brown, B. J. Callahan, A. M. Caraballo-Rodríguez, J. Chase, E. K. Cope, R. Da Silva, C. Diener, P. C. Dorrestein, G. V. Douglas, D. M. Durall, C. Duvallet, C. F. Edwardson, M. Ernst, M. Estaki, J. Fouquier, J. M. Gauglitz, S. M. Gibbons, D. L. Gibson, A. Gonzalez, K. Gorlick, J. Guo, B. Hillmann, S. Holmes, H. Holste, C. Huttenhower, G. A. Huttley, S. Janssen, A. K. Jarmusch, L. Jiang, B. D. Kaehler, K. B. Kang, C. R. Keefe, P. Keim, S. T. Kelley, D. Knights, I. Koester, T. Kosciolek, J. Kreps, M. G. I. Langille, J. Lee, R. Ley, Y.-X. Liu, E. Loftfield, C. Lozupone, M. Maher, C. Marotz, B. D. Martin, D. McDonald, L. J. McIver, A. V. Melnik, J. L. Metcalf, S. C. Morgan, J. T. Morton, A. T. Naimey, J. A. Navas-Molina, L. F. Nothias, S. B. Orchanian, T. Pearson, S. L. Peoples, D. Petras, M. L. Preuss, E. Pruesse, L. B. Rasmussen, A. Rivers, M. S. Robeson, P. Rosenthal, N. Segata, M. Schaffer, A Shiffer, R. Sinha, S. J. Song, J. R. Spear, A. D. Swafford, L. R. Thompson, P. J. Torres, P. Trinh, A. Tripathi, P. J. Turnbaugh, S. Ul-Hasan, J. J. J. van der Hofft, F. Vargas, Y. Vázquez-Baeza, E. Vogtmann, M. von Hippel, W. Walters, Y. Wan, M. Wang, J. Warren, K. C. Weber, C. H. D. Williamson, A. D. Willis, Z. Z. Xu, J. R. Zaneveld, Y. Zhang, Q. Zhu, R. Knight, and J. G. Caporaso. 2019. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat. Biotechnol. 37:852–857. doi: https://doi.org/10.1038/s41587-019-0209-9
Buege, J. A., and S. D. Aust. 1978. Microsomal lipid peroxidation. In: S. Fleischer and L. Packer, editors, Methods in enzymology, vol. 52. Academic Press. p. 302–310. doi: https://doi.org/10.1016/S0076-6879(78)52032-6
Burcham, Z. M., A. D. Belk, B. B. McGivern, A. Bouslimani, P. Ghadermazi, C. Martino, L. Shenhav, A. R. Zhang, P. Shi, A. Emmons, H. L. Deel, Z. Z. Xu, V. Nieciecki, Q. Zhu, M. Shaffer, M. Panitchpakdi, K. C. Weldon, K. Cantrell, A. Ben-Hur, S. C. Reed, G. C. Humphry, G. Ackermann, D. McDonald, S. H. J. Chan, M. Connor, D. Boyd, J. Smith, J. M. S. Watson, G. Vidoli, D. Steadman, A. M. Lynne, S. Bucheli, P. C. Dorrestein, K. C. Wrighton, D. O. Carter, R. Knight, and J. L. Metcalf. 2024. A conserved interdomain microbial network underpins cadaver decomposition despite environmental variables. Nat. Microbiol. 9:595–613. doi: https://doi.org/10.1038/s41564-023-01580-y
Callahan, B. J., P. J. McMurdie, M. J. Rosen, A. W. Han, A. J. A. Johnson, and S. P. Holmes. 2016. DADA2: high-resolution sample inference from Illumina amplicon data. Nat. Methods. 2016; 13:581–583. doi: https://doi.org/10.1038/nmeth.3869
Chen, J., J. Zhang, N. Wang, B. Xiao, X. Sun, J. Li, K. Zhong, L. Yang, X. Pang, F. Huang, and A. Chen. 2024. Critical review and recent advances of emerging real-time and non-destructive strategies for meat spoilage monitoring. Food Chem. 445:138755. doi: https://doi.org/10.1016/j.foodchem.2024.138755
Chen, X., L. J. Bauermeister, G. N. Hill, M. Singh, S. F. Bilgili, and S. R. Mckee. 2014. Efficacy of various antimicrobials on reduction of Salmonella and Campylobacter and quality attributes of ground chicken obtained from poultry parts treated in a postchill decontamination tank. J. Food Prot. 77:1882–1888. doi: https://doi.org/10.4315/0362-028X.JFP-14-114
Faith, D. P. 1992. Conservation evaluation and phylogenetic diversity. Biol. Conserv. 61:1–10. doi: https://doi.org/10.1016/0006-3207(92)91201-3
Fang, J., L. Feng, H. Lu, and J. Zhu. 2022. Metabolomics reveals spoilage characteristics and interaction of Pseudomonas lundensis and Brochothrix thermosphacta in refrigerated beef. Food Res. Int. 156:111139. doi: https://doi.org/10.1016/j.foodres.2022.111139
FAO. 2025. Technical platform on the measurement and reduction of food loss and waste (TPFLW): food loss and waste database. Food and Agriculture Organization of the United Nations. https://fao.org/platform-food-loss-waste/flw-data/en/. (Accessed March 2025)https://fao.org/platform-food-loss-waste/flw-data/en/
Fidan, H., T. Esatbeyoglu, V. Simat, M. Trif, G. Tabanelli, T. Kostka, C. Montanari, S. A. Ibrahim, and F. Özogul. 2022. Recent developments of lactic acid bacteria and their metabolites on foodborne pathogens and spoilage bacteria: facts and gaps. Food Biosci. 47:101741. doi: https://doi.org/10.1016/j.fbio.2022.101741
Gram, L., L. Ravn, M. Rasch, J. B. Bruhn, A. B. Christensen, and M. Givskov. 2002. Food spoilage—interactions between food spoilage bacteria. Int. J. Food Microbiol. 78:79–97. doi: https://doi.org/10.1016/S0168-1605(02)00233-7
Gray, S. 2019. Expiration, use-by and sell-by dates: what do they really mean? UCONN Extension. https://publications.extension.uconn.edu/2019/10/30/expiration-use-by-and-sell-by-dates-what-do-they-really-mean/. (Accessed June 2025)https://publications.extension.uconn.edu/2019/10/30/expiration-use-by-and-sell-by-dates-what-do-they-really-mean/
Hernández-García, E. M. Vargas, and S. Torres-Giner. 2022. Quality and shelf-life stability of pork meat fillets packaged in multilayer polylactide films. Foods. 11:426. doi: https://doi.org/10.3390/foods11030426
Hu, Y., M. Xu, X. Gao, and Z. Zhang. 2024. Influence of chitosan/lycopene on myoglobin and meat quality of beef during storage. Coatings. 14:1445. doi: https://doi.org/10.3390/coatings14111445
Hultman, J., R. Rahkila, J. Ali, J. Rousu, and K. J. Björkroth. 2015. Meat processing plant microbiome and contamination patterns of cold-tolerant bacteria causing food safety and spoilage risks in the manufacture of vacuum-packaged cooked sausages. Appl. Environ. Microbiol. 81:7088–7097. doi: https://doi.org/10.1128/AEM.02228-15
Hung, Y.-M., W.-N. Lyu, M.-L. Tsai, C.-L. Liu, L.-C. Lai, M.-H. Tsai, and E. Y. Chuang. 2022. To compare the performance of prokaryotic taxonomy classifiers using curated 16S full-length rRNA sequences. Comput. Biol. Med. 145:105416. doi: https://doi.org/10.1016/j.compbiomed.2022.105416
Hwang, B. K., H. Choi, S. H. Choi, and B.-S. Kim. 2020. Analysis of microbiota structure and potential functions influencing spoilage of fresh beef meat. Front. Microbiol. 11:1657. doi: https://doi.org/10.3389/fmicb.2020.01657
Jafarzadeh, S., Z. Yildiz, P. Yildiz, P. Strachowski, M. Forough, Y. Esmaeili, M. Naebe, and M. Abdollahi. 2024. Advanced technologies in biodegradable packaging using intelligent sensing to fight food waste. Int. J. Biol. Macromol. 261:129647. doi: https://doi.org/10.1016/j.ijbiomac.2024.129647
Jayan, H., W. Min, and Z. Guo. 2025. Applications of artificial intelligence in food industry. Foods. 14:1241. doi: https://doi.org/10.3390/foods14071241
King, D. A., M. C. Hunt, S. Barbut, J. R. Claus, D. P. Cornforth, P. Joseph, Y. H. B. Kim, G. Lindahl, R. A. Mancini, M. N. Nair, K. J. Merok, A. Milkowski, A. Mohan, F. Pohlman, R. Ramanathan, C. R. Raines, M. Seyfert, O. Sørheim, S. P. Suman, and M. Weber. 2023. American Meat Science Association guidelines for meat color measurement. Meat and Muscle Biology. 6:12473. doi: https://doi.org/10.22175/mmb.12473
Lenth, R. V., J. Piaskowski, B. Banfai, B. Bolker, P. Buerkner, I. Giné-Vázquez, M. Hervé, M. Jung, J. Love, F. Miguez, H. Riebl, and H. Singmann. 2025. Emmeans: estimated marginal means, aka least-squares means. 2.0.1. doi: https://doi.org/10.32614/CRAN.package.emmeans
Li, X., Y. Wang, H. Ma, Q. Cao, Q. Zhang, Y. Li, F. Ma, D. Xie, B. Zhang, G. Liu, J. Cai, and Y. Zhao. 2025. The effect of mixed natural preservatives on microbial community and Tan mutton quality. J. Food Meas. Charact. 19:2942–2956. doi: https://doi.org/10.1007/s11694-025-03157-1
Lin, H., and S. D. Peddada. 2020. Analysis of compositions of microbiomes with bias correction. Nat. Commun. 11:3514. doi: https://doi.org/10.1038/s41467-020-17041-7
Lozupone, C., and R. Knight. 2005. UniFrac: a new phylogenetic method for comparing microbial communities. Appl. Environ. Microbiol. 71:8228–8235. doi: https://doi.org/10.1128/AEM.71.12.8228-8235.2005
Lozupone, C., M. E. Lladser, D. Knights, J. Stombaugh, and R. Knight. 2011. UniFrac: an effective distance metric for microbial community comparison. ISME J. 5:169–172. doi: https://doi.org/10.1038/ismej.2010.133
Mamat, M., F. Wong, H. T. Yew, and J. A. Dargham. 2024. Estimating shelf life of packed fresh milk using odor and machine learning: a feasibility study. In: H. T. Yew, M. Mamat, J. A. Dargham, C. S. Kheau, and E. G. Moung, editors, Internet of things and artificial intelligence for smart environments. Springer Nature, Singapore. p.143–165. doi: https://doi.org/10.1007/978-981-97-1432-2_9
Martin, J. N., J. C. Brooks, T. A. Brooks, J. F. Legako, J. D. Starkey, S. P. Jackson, and M. F. Miller. 2013. Storage length, storage temperature, and lean formulation influence the shelf-life and stability of traditionally packaged ground beef. Meat Sci. 95:495–502. doi: https://doi.org/10.1016/j.meatsci.2013.05.032
Metcalf, J. L., Z. Z. Xu, S. Weiss, S. Lax, W. van Treuren, E. R. Hyde, S. J. Song, A. Amir, P. Larsen, N. Sangwan, D. Haarmann, G. C. Humphrey, G. Ackermann, L. R. Thompson, C. Lauber, A. Bibat, C. Nicholas, M. J. Gebert, J. F. Petrosino, S. C. Reed, J. A. Gilbert, A. M. Lynne, S. R. Bucheli, D. O. Carter, and R. Knight. 2015. Microbial community assembly and metabolic function during mammalian corpse decomposition. Science. 351:158–162. doi: https://doi.org/10.1126/science.aad2646
Odeyemi, O. A., O. O. Alegbeleye, M. Strateva, and D. Stratev. 2020. Understanding spoilage microbial community and spoilage mechanisms in foods of animal origin. Compr. Rev. Food Sci. Food Saf. 19:311–331. doi: https://doi.org/10.1111/1541-4337.12526
Palanisamy, Y., V. Kadirvel, N. D. Ganesan. 2024. Recent technological advances in food packaging: sensors, automation, and application. Sustainable Food Technol. 3:161–180. doi: https://doi.org/10.1039/d4fb00296b
Papoutsoglou, G., S. Tarazona, M. B. Lopes, T. Klammsteiner, E. Ibrahimi, J. Eckenberger, P. Novielli, A. Tonda, A. Simeon, R. Shigdel, S. Béreux, G. Vitali, S. Tangaro, L. Lahti, A. Temko, M. J. Claesson, and M. Berland. 2023. Machine learning approaches in microbiome research: challenges and best practices. Front. Microbiol. 14:1261889. doi: https://doi.org/10.3389/fmicb.2023.1261889
Pedregosa, F., G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weis, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, and É. Duchesnay. 2011. Scikit-learn: machine learning in python. J. Mach. Learn. Res. 12:2825–2830.
Pellissery, A. J., P. G. Vinayamohan, M. A. R. Amalaradjou, and K. Venkitanarayanan. 2020. Spoilage bacteria and meat quality. In: A. K. Biswas and P. K. Mandal, editors, Meat quality analysis: advanced evaluation methods, techniques, and technologies. Academic Press. p. 307–334. doi: https://doi.org/10.1016/B978-0-12-819233-7.00017-3
Quast, C., E. Pruesse, P. Yilmaz, J. Gerken, T. Schweer, P. Yarza, J. Peplies, and F. O. Glöckner. 2013. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 41:D590–D596. doi: https://doi.org/10.1093/nar/gks1219
R Core Team. 2022. The R project for statistical computing. R Foundation for Statistical Computing, Vienna. https://R-project.org. (Accessed January 15, 2023)https://R-project.org
Snyder, A. B., N. Martin, and M. Wiedmann. 2024. Microbial food spoilage: impact, causative agents and control strategies. Nat. Rev. Microbiol. 22:528–542. doi: https://doi.org/10.1038/s41579-024-01037-x
Stanborough, T., N. Fegan, S. M. Powell, T. Singh, M. Tamplin, and P. S. Chandry. 2018. Genomic and metabolic characterization of spoilage-associated Pseudomonas species. Int. J. Food Microbiol. 268:61–72. doi: https://doi.org/10.1016/j.ijfoodmicro.2018.01.005
Taormina, P. J. 2021. Microbial growth and spoilage. In: P. J. Taormina and M. D. Hardin, editors, Food safety and quality-based shelf life of perishable foods. Springer, Cham, Switzerland. p. 41–69 doi: https://doi.org/10.1007/978-3-030-54375-4_3
Tarlak, F. 2023. The use of predictive microbiology for the prediction of the shelf life of food products. Foods. 12:4461. doi: https://doi.org/10.3390/foods12244461
Wickham, H., W. Chang, L. Henry, T. L. Pederson, K. Takahashi, C. Wilke, K. Woo, H. Yutani, D. Dunnington, and T. van den Brand. 2016. Ggplot2: elegant graphics for data analysis. Springer-Verlag, New York. https://ggplot2.tidyverse.org. (Accessed June 2025)https://ggplot2.tidyverse.org
Wickham, H., M. Averick, J. Bryan, W. Chang, L. D. McGowan, R. François, G. Grolemund, A. Hayes, L. Henry, J. Hester, M. Kuhn, T. L. Pederson, E. Miller, S. M. Bache, K. Müller, J. Ooms, D. Robinson, D. P. Seidel, V. Spinu, K. Takahashi, D. Vaughan, C. Wilke, K. Woo, and H. Yutani. 2019. Welcome to the tidyverse. J. Open Source Software. 4:1686. doi: https://doi.org/10.21105/joss.01686
Wickramasinghe, N. N., J. Ravensdale, R. Coorey, S. P. Chandry, and G. A. Dykes. 2019. The predominance of psychrotrophic pseudomonads on aerobically stored chilled red meat. Compr. Rev. Food Sci. Food Saf. 18:1622–1635. doi: https://doi.org/10.1111/1541-4337.12483
Xia, X.-x., C.-x. Li, and H.-r. Guo. 2025. Association between oral microbiome diversity and chronic obstructive pulmonary disease in the US Population. J. Transl. Med. 23:557. doi: https://doi.org/10.1186/s12967-025-06553-9
Xia, L., M. Qian, F. Cheng, Y. Wang, J. Han, Y. Xu, K. Zhang, J. Tian, and Y. Jin. 2023. The effect of lactic acid bacteria on lipid metabolism and flavor of fermented sausages. Food Biosci. 56:103172. doi: https://doi.org/10.1016/j.fbio.2023.103172
Xia, Y., and J. Sun. 2023. Beta diversity metrics and ordination. In: Y. Xia and J. Sun, editors, Bioinformatic and statistical analysis of microbiome data: from raw sequences to advanced modeling with QIIME 2 and R. Springer, Cham, Switzerland. p. 335–395. doi: https://doi.org/10.1007/978-3-031-21391-5_10
Xiao, L., L. Cui, M. Lapu, T. Bai, X. Guo, D. Liu, M. Liu, and X. Wang. 2025. The structure, assembly processes of microbial communities and their effects on the quality of goat meat during chilled storage (4 °C). Foods. 14:1653. doi: https://doi.org/10.3390/foods14091653
Xie, V., and R. Bagchi. 2024. How duration of storage affects food waste behavior. Journal of Consumer Psychology. 34:570–587. doi: https://doi.org/10.1002/jcpy.1389
Xu, Z. S., J. Hettinger, A. Athey, X. Yang, and M. G. Gänzle. 2025. Control of meat spoilage with ozone nano-bubbles: insights from laboratory model systems and commercial scale treatments. Int. J. Food Microbiol. 433:111128. doi: https://doi.org/10.1016/j.ijfoodmicro.2025.111128
Zhang, G., and Y. Lu. 2012. Bias-corrected random forests in regression. J. Appl. Stat. 39:151–160. doi: https://doi.org/10.1080/02664763.2011.578621
Supplemental Material
Ground beef spoilage shown in images throughout the 14-d shelf-life study.
(Left) β diversity analysis comparing distances of microbial communities associated with nonspoiled samples to spoiled samples, (right) and spoiled samples to nonspoiled samples.
Instrumental color values for L* (lightness), a* (redness), and b* (yellowness), demonstrating changes in the values throughout the 14-d experimental period.
| Day | Average L* | Average a* | Average b* |
|---|---|---|---|
| 0 | 55.227 | 34.810 | 26.117 |
| 1 | 55.996 | 33.544 | 25.130 |
| 2 | 52.380 | 19.574 | 16.675 |
| 3 | 52.955 | 16.449 | 16.326 |
| 4 | 52.803 | 11.214 | 14.751 |
| 5 | 53.652 | 9.644 | 14.943 |
| 6 | 52.718 | 10.550 | 14.878 |
| 7 | 51.756 | 11.675 | 14.420 |
| 8 | 50.402 | 13.734 | 15.514 |
| 9 | 45.926 | 18.610 | 19.411 |
| 10 | 49.971 | 16.968 | 16.247 |
| 11 | 47.602 | 17.399 | 16.209 |
| 12 | 49.859 | 17.004 | 15.261 |
| 13 | 47.524 | 18.174 | 15.622 |
| 14 | 47.616 | 17.263 | 14.251 |
Ground beef compositional analysis values obtained through FoodScan™, demonstrating little change throughout the experimental period.
| Day | Protein | Moisture | Fat |
|---|---|---|---|
| 0 | 19.11 | 68.28 | 10.38 |
| 1 | 18.67 | 68.34 | 10.42 |
| 2 | 18.69 | 68.10 | 10.48 |
| 3 | 18.52 | 68.07 | 10.36 |
| 4 | 18.35 | 67.46 | 10.42 |
| 5 | 18.40 | 67.77 | 10.59 |
| 6 | 18.79 | 67.79 | 10.39 |
| 7 | 19.38 | 68.03 | 10.36 |
| 8 | 19.61 | 67.92 | 10.61 |
| 9 | 19.88 | 68.12 | 10.55 |
| 10 | 19.86 | 68.16 | 10.47 |
| 11 | 19.95 | 68.07 | 10.76 |
| 12 | 20.14 | 68.36 | 10.49 |
| 13 | 20.25 | 67.87 | 10.83 |
| 14 | 20.21 | 68.31 | 10.89 |