J. Dairy Sci.

107:4259–4276
https://doi.org/10.3168/jds.2023-24364
© 2024, The Authors. Published by Elsevier Inc. on behalf of the American Dairy Science Association®.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Evaluation of kefir grain microbiota, grain viability, and bioactivity

from fermenting dairy processing by-products
Chloe J. McGovern, Brianda D. González-Orozco, and Rafael Jiménez-Flores*
Department of Food Science and Technology, The Ohio State University, Columbus, OH 43210

ABSTRACT lowest Trolox equivalence concentration in the ABTS as-

say. Sweet whey and supplemented milk fat sweet whey
Four dairy foods processing by-products (acid whey had upregulation of Cldn-1 and Ocln-1 gene expression,
permeate [AWP], buttermilk [BM], sweet whey permeate which correspond with a significant increase in transepi-
[SWP], and sweet whey permeate with added milk fat thelial electrical resistance.
globule ingredient [SWP+MFGM]) were fermented for Key words: kefir grain microbiota, metagenomics, dairy
4 wk and compared with traditional kefir milks for pro- by-products, novel dairy beverage, kefir fermentation
duction of novel kefir-like dairy products. Sweet whey
permeates and SWP supplemented with 1.5% milk fat
INTRODUCTION
globule membrane (MFGM) proved to be the most vi-
able by-products for kefir grain fermentation, exhibiting Kefir is increasing in popularity due to increased
diverse abundance of traditional kefir microorganisms consumer interest in gut health and probiotic products,
and positive indicators of bioactive properties. Grain vi- as well as research focusing on the intersectionality at
ability was assessed with shotgun metagenomics, texture the gut-brain axis (Pereira et al., 2021). Commercial
profile analysis, live cell counts, and scanning electron kefir can be purchased at the grocery store and closely
microscopy. Assessed bioactivities of the kefir-like resembles a drinkable yogurt fermented with common
products included antibacterial, antioxidant, potential probiotics, or kefir can be produced at home from kefir
anticancerogenic properties, and membrane barrier ef- grains for a more traditional product. Traditional kefir is
fects on human colorectal adenocarcinoma Caco-2 cells. a fermented milk beverage made by the addition of kefir
All kefir grains were most abundant in Lactobacillus grains (5%–10% wt/vol) into milk and fermenting for 12
kefiranofaciens when analyzed with shotgun metage- to 48 h. The grains are strained out of the fermented milk
nomics. When analyzed with live cell counts on selec- and then the final fermented product, kefir, is ready to
tive media, AWP kefir-like product had no countable consume (Rosa et al., 2017). Kefir grains are a complex
Lactococcus spp., indicating suboptimal conditions for consortium of bacteria and yeast that have a symbiotic
kefir grain microbiota survival and application for fer- relationship. The kefir grain microorganisms produce a
mented dairy starter culture bacterium. Live cell counts variety of metabolites known to give kefir its charac-
were affirmed with kefir grain surface scanning electron teristic sour and alcoholic flavor and contribute to the
microscopy images. The SWP treatment had the most positively associated health benefits. Some cited positive
adhesive kefir grain surface, and SWP+MFGM had the bioactive properties of kefir are antioxidant properties
largest exopolysaccharide yield from grain extraction. (Liu et al., 2005; Erdogan et al., 2019), antibacterial
All kefir and kefir-like products were able to achieve a properties (Liu et al., 2022), and positive gut microbiome
6-log reduction against Listeria innocua and Escherichia results (Tung et al., 2018; Bellikci-Koyu et al., 2019; Er-
coli. Traditional milk kefirs had the highest antioxidant dogan et al., 2019), which are attributed to the complex
capacity for 2,2-diphenyl-1-picrylhydrazyl (DPPH) and microbial diversity of kefir grain microorganisms and
the 2,2′-azino-bis (3-ethylbenzothiazoline-6-sulfonic their metabolites.
acid; ABTS) assay. The AWP formulation had a signifi- The microbial composition of kefir grains is depen-
cantly higher DPPH antioxidant activity compared with dent on region, mother grain, milk type, and fermenta-
the other kefir and kefir-like products, and SWP had the tion conditions (Avila-Reyes et al., 2022). Although a
kefir grain has yet to be produced from a laboratory by
judiciously combining microorganisms, research indi-
Received October 27, 2023. cates common genera and species that make up most
Accepted January 12, 2024.
*Corresponding author: jimenez-flores.1@osu.edu of the kefir grain relative abundance. Genera include

The list of standard abbreviations for JDS is available at adsa.org/jds-abbreviations-24. Nonstandard abbreviations are available in the Notes.

4259
 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4260
Lactobacillus spp., Lactococcus spp., Acetobacter spp., MATERIALS AND METHODS
and “contaminants” such as Pseudomonas spp. and En-
terococcus spp. (Gao et al., 2013; Nalbantoglu et al., By-Product Preparation
2014; González-Orozco et al., 2023), among others.
Ultra-high temperature whole milk (WM; Kroger, Cin-
Lactobacillus kefiranofaciens is a predominant bacte-
cinnati, OH), UHT skim milk (SM; Natrel, Quebec, Can-
rial species (98.3%) found in kefir grain when metage-
ada), BM, AWP, SWP, and SWP supplemented with 1.5%
nomically analyzed, but not a majority species in kefir
milk fat globule membrane ingredient (SWP+MFGM)
itself (González-Orozco et al., 2023). The species L.
were used as fermentation substrates for kefir grains. The
kefiranofaciens produces an exopolysaccharide (EPS)
BM and acid whey were produced in The Ohio State Uni-
known as kefiran, an integral part in the structure of
versity Dairy Processing Pilot Plant (Department of Food
the kefir grain and attributed with technological and
Science and Technology, The Ohio State University, Co-
prebiotic applications (Maeda et al., 2004; Piermaria et
lumbus, OH). Sweet whey was received from Guggisberg
al., 2008; Wang et al., 2008).
Cheese (Guggisberg Cheese Inc., Millersburg, OH). The
Whole cow milk and skim milk are the most common
acid and sweet whey were filtered using a benchtop 10-
substrates used for commercial kefir production. How-
kDa membrane (Pellicon 3 Cassettes, Millipore Sigma,
ever, traditional kefir has been made with other milks
Burlington, MA) ultrafiltration system (Ultrafiltration
from goats or buffalo (Tomar et al., 2020; Azizkhani
System, Millipore Sigma). The MFGM (Lactprodan
et al., 2021). Previous reports have explored the fer-
MFGM-10, Arla, Denmark) was added at 1.5% wt/vol to
mentation of whey with kefir grains to form a kefir-like
the SWP. Permeates and buttermilk were batch pasteur-
beverage. Based on the volatiles in kefir, whey-based
ized after filtration at 62°C for 30 min.
kefir-like beverages have been found to be similar in
ethanol, lactic acid, and other volatiles associated with
Proximate Macromolecule Composition
flavor and sensory of kefir in whey and deproteinated
“raw” whey (Magalhães et al., 2011). Another dairy By-products were analyzed for their overall composi-
processing by-product is buttermilk from the butter tion. Total solids and moisture were analyzed in a CEM
churn. Buttermilk is high in milk phospholipids (MPL) Smart 6 (CEM Corporation). Protein concentration
and milk fat components (Sodini et al., 2006). The MPL was evaluated using CEM Sprint (CEM Corporation).
and milk fat components are derivatives of the milk fat Fat was determined using the CEM Oracle (CEM Cor-
globule membrane (MFGM), which has been demon- poration). Ash content was determined with the CEM
strated to enhance survival of probiotic microorganisms Phoenix (CEM Corporation). Carbohydrate amount was
through digestion in addition to other health benefits of determined by difference.
milk fat and probiotics (Kosmerl et al., 2021), making
milk an ideal medium for probiotic-focused foods such Kefir Grain Revitalization and Adaptation
as kefir. to By-Product Media
Fermentation is a complex metabolic process that
transforms the simplest of ingredients into a delicacy, Kefir grains were purchased from Mr. and Mrs. Kefir
such as wine or cheese, by the action of microorgan- Company. To allow for adaptation of the grain micro-
isms. The metabolic process of homo- and heterofer- biota to each by-product medium, the grains were split
mentative bacteria found in a kefir grain microbiota has into 4.5-g pieces and revitalized at 10% wt/vol in each
the potential to transform dairy processing by-products medium for 72 h at 25°C undisturbed. Next, grains were
into new dairy beverages. Acid whey permeate (AWP) filtered with a sieve and added back into a sterile bottle
from cottage cheese production, sweet whey permeates with 10% wt/vol of the medium. The fermented prod-
(SWP) from Swiss cheese production, and buttermilk ucts were filtered every 48 h at (10% wt/vol) for a total
(BM) from a continuous churn will be evaluated for the of a 7-d revitalization and adaptation period. Following
survival of kefir microorganisms, overall grain viability, the medium change on d 7 (D7), the 4-wk fermentation
and final kefir-like fermented product bioactivity. The cycle began, with D7 of revitalization equating to wk 0
objective of this research is to evaluate the application (W0) of fermentation. During the fermentation, grain
of dairy processing by-products for fermented dairy weight and pH were taken every 48 h at the changing
starter cultures and the development of novel fermented of media.
dairy products through fermentation with kefir grains.
The microbial complexity of kefir grains makes them DNA Extraction for Shotgun Metagenomics
an ideal candidate for transforming dairy processing
by-products and waste into a potential novel fermented To evaluate the effects of the fermentation media in
dairy product. the composition of the kefir grain microbiota, shotgun
Journal of Dairy Science Vol. 107 No. 7, 2024
 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4261

metagenomics were performed with the assistance of Di- Scanning Electron Microscopy of Kefir Grains
versigen Inc. (Minneapolis, MN). Using 2 sterile transfer
sticks, aliquots of each kefir grain were plated into Pow- Surface imaging of the kefir grains in different me-
erBead Pro lysis plates (Qiagen, Hilden, Germany) at 1 dia may provide insight to the microbial population and
g, 0.5 g, and 0.3 g. Samples were resuspended in lysis structure of the kefir grain shown in shotgun metage-
buffer CD1 (800 µL) with 20 µL of Proteinase K (20 mg/ nomic analysis and live cell counts. Grain samples at
mL) and incubated for 2 h at 50°C. Following incubation, W4 were filtered from their media with a sieve, rinsed in
genomic DNA (gDNA) was extracted following the steps PBS, and submerged in PBS to ensure total cleaning of
of the Diversigen Inc. protocol (SOP-41). Quality of the grains and neutralization of the grain pH for the fixation
extracted gDNA was determined by the PicoGreen quan- adhesive. Grains were added to fixative solution (3% glu-
tification (SOP-17) method. Extractions were performed taraldehyde, 2% paraformaldehyde in 0.1 M potassium
in triplicate. phosphate buffer [PPB], pH 7.2), vacuum infiltrated, and
placed on a rocker overnight at 25°C. Grains were washed
for 15 min in PPB, then rinsed 3 times in MilliQ water
Library Preparation and Sequencing of Kefir
(18.2 MW·cm). The grains were dehydrated in 15-min
Grain DNA
intervals in 25%, 50%, 70%, and 90% ethanol, followed
by 3 final rinses in 100% ethanol for 15 min each. Grains
Creation of the DNA library and shallow shotgun
were put on a rotator and spray coated in platinum to a
sequencing were performed by Diversigen Inc. with
thickness of 5 nm. Grains were imaged on the Schottky
the standard BoosterShot workflow. Libraries were se-
Field Emission Scanning Electron Microscope (Hitachi
quenced at a target depth of 2M with paired end reads
High Technologies America, Inc., Greenville, SC) at The
on a NovaSeq 6000 using a 2× 150-bp flow cell. Reads
Ohio State Molecular and Cellular Imaging Center in
were run in CCMetagen 1.3 software (Ondov et al., 2011;
Wooster, Ohio.
Clausen et al., 2018) from the Center for Genomic Epide-
miology (National Food Institute, Technical University
of Denmark, Lyngby, Denmark). Kefir Grain TPA
The W4 grain samples were analyzed for adhesiveness
Kefir Grain Bioavailability Assessed by Live (N × s) with the TA.XT Plus Texture Analyzer (Stable
Cell Counts Micro Systems, Godalming, UK) with a 6.35-mm diam-
eter and a 35-mm-long acrylamide cylinder head. Surface
Metagenomic analysis detects DNA from both viable adhesiveness is an indicator for stickiness or tackiness
and non-viable microorganisms without distinction. analysis on foods that are commonly associated with
Therefore, estimating viable cell quantities throughout polysaccharides. The EPS found on the surface of kefir
the fermentation period aids in a more comprehensive grains have an adhesive property that can be estimated
interpretation of metagenomic data. Live cell counts with rheology to indicate EPS production from the kefir
were determined at W0, wk 2 (W2), and wk 4 (W4) grain bacteria.
from each fermented product. Each fermented product
was serially diluted in sterile peptone water (Sigma- Extraction of EPS from Kefir Grain
Aldrich) and plated in duplicate on selective media.
Lactobacillus spp. were cultured on lactobacillus- An important structural component of kefir grains,
selective De Man, Rogosa, Sharpe (MRS) medium (BD EPS can play a role in the establishment, survival, and
Difco, Fisher Scientific, Waltham, MA), with 7.5% development of interactions between the kefir grain mi-
agar (Sigma-Aldrich) added, and supplemented with croorganisms. To extract EPS, kefir grains were boiled
20 mL/L of agar of a 0.5 mg/mL cycloheximide (Fisher at 100°C for 60 min to lyse cells (Rimada and Abraham,
Scientific) solution in 96% ethanol. Lactococcus spp. 2001), and, after reaching room temperature, the boiled
were counted on M17 agar (Sigma-Aldrich). Eukary- grain was centrifuged (5804 R, Eppendorf, Hamburg,
otes were cultured on Rose Bengal chloramphenicol Germany) in 50-mL falcon tubes at 7,546 × g for 25 min.
agar (RBCA, Oxoid, Fisher Scientific) supplemented Supernatant was transferred to 50-mL falcon tubes (10
with 1 mL/L of 25 mg/mL solution chloramphenicol in mL), and 30 mL of 200-proof cold (4°C) ethanol was
96% ethanol (MP Biomedicals, Santa Ana, CA). The added. Tubes were stored at −20°C for 24 h (Lim et
M17 plates were incubated at 30°C, RBCA plates at al., 2017). Afterward, tubes were centrifuged at 13,416
25°C, and MRS plates at 37°C in anaerobic chambers, × g for 20 min, and the resulting supernatant was dis-
all for 48 h. carded. To resuspend the pellet, 20 mL of MilliQ water

Journal of Dairy Science Vol. 107 No. 7, 2024

 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4262
(18.2 MW·cm) was added to each tube and mixed until a 0.1 mM solution of DPPH was prepared in methanol
complete resuspension of the pellet. The hydrated sus- (Rahman et al., 2015; Baliyan et al., 2022). In a 96-
pension was frozen at −80°C and lyophilized. Weight of well plate, 120 µL of the DPPH solution was added to
EPS post-lyophilization was measured to determine yield each well with 80 µL of the supernatants. The plate was
from 5 g of grain. shaken and incubated in the dark for 30 min at 25°C.
The plate was then centrifuged at 143 × g for 10 min
Colorimetric Assays to Estimate Total EPS from at 25°C. Supernatants were transferred to a new 96-well
Crude Extract plate, and absorbance was recorded at 517 nm. Methanol
was used as the control sample, and different concen-
Total EPS in µg/mL was determined by subtracting the trations of Trolox were used to create a standard curve.
total sugar content assessed by the phenol sulfuric assay Samples were read in triplicate, and the percentage of
method (DuBois et al., 1956) from the reducing estimated DPPH radical scavenging activity was calculated using
by the DNS (dinitrosalicylic acid) method (Gonçalves et the following equation:
al., 2010) with 1:1 ratio of glucose and galactose glucose
as a standard (Enikeev, 2012). Subtracting the estimated % Radical scavenging activity
total sugar from the estimated reducing sugar content
Abs control − Abs sample 
 
should provide an estimated quantity of nonreducing =   ×100,
EPS (Ruijssenaars et al., 2000).

 Abs control 

Bioactivity of Kefir and Kefir-Like Final where Abscontrol is the absorbance of the methanol blank,
Fermented Products and Abssample is the absorbance of the supernatants.
For the ABTS method, a 7 mM solution of ABTS
Antioxidant activity, antibacterial activity, and mainte- (Sigma Aldrich) was prepared in MilliQ water (18.2
nance of the intestinal membrane barrier are some of the MW·cm). A second solution of 254 mM ammonium
bioactive properties commonly associated with kefir. In persulfate (APS; Sigma Aldrich) was prepared in Mil-
some bioactivity assays, kefir supernatant is used instead liQ water (18.2 MW·cm). A final solution of ABTS+ was
of kefir due to spectrophotometry limitations. Superna- prepared to a final APS concentration of 2.54 mM (Re et
tant of the kefir and kefir-like products was collected by al., 1999). The solution was incubated at 25°C overnight
centrifugation at 4,000 × g for 10 min at 4°C. in the dark. Following incubation, 190 µL of ABTS+ so-
lution was added to a well with 10 µL of the kefir-like
Antibacterial Activity of Kefir and Kefir-Like supernatants and Trolox standards. MilliQ water (18.2
Fermented Products MW·cm) was used as the control. Samples were done
in triplicate. The average absorbance of the samples was
Listeria innocua ATCC 25922 and Escherichia coli used to calculate the percentage of decolorization with
ATCC 25922 were used as indicator organisms to assess the following equation:
the antibacterial activity of kefir and kefir-like fermented
products using optical density (OD) and plate counts to % Radical scavenging activity
determine log reduction of surrogate pathogens over Abs control − Abs sample 
 
time. The kefir and kefir-like products’ supernatants were =   ×100,
filtered with a 0.2-µm filter before inoculation. Surrogate

 Abs control 
pathogen and supernatant ratio were 1:1, resulting in a
mixture with 100 µL of supernatant and 100 µL of E. where Abscontrol is the absorbance of the water blank, and
coli and L. innocua (~7 log cfu/mL). Plating was done Abssample is the absorbance of the kefir supernatant and
on TSA at 0, 6, 12, 18, and 24 h. Plates were incubated Trolox standards. A standard curve of Trolox at different
for 24 h at 37°C every 4 h to correlate OD, time, and log concentrations was generated, to report results as micro-
cfu/mL. molar Trolox.

Antioxidant Activity of Kefir and Kefir-Like Growth and Preparation of Caco-2 Cells
Fermented Products
The Caco-2 cell line (HTB37, American Type Culture
The 2,2-diphenyl-1-picrylhydrazyl (DPPH) and the Collection, Manassas, VA) from human colorectal ad-
2,2′-azino-bis (3-ethylbenzothiazoline-6-sulfonic acid) enocarcinoma was used to model intestinal epithelium.
(ABTS) methods were used to determine the antioxidant Cells were grown in T-75 cm2 flasks (Corning, Corning,
activity of the kefir and kefir-like supernatants. Briefly, NY) with complete Dulbecco’s modified Eagle’s medium
Journal of Dairy Science Vol. 107 No. 7, 2024
 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4263

(cDMEM) supplemented with 10% heat-inactivated fetal (Millipore Sigma) stored and washed with sterile DPBS
bovine serum, 1% penicillin-streptomycin (100 units/ (Hubatsch et al., 2007). Experiments were performed on
mL penicillin and 100 units/mL streptomycin), 1% non- 24-well membrane inserts (0.4-μm PET tracked-etched
essential AA (100×), and 1% 200 mM L-glutamine in a membrane, Corning), seeded apically with 0.5 mL of
humidified atmosphere of air and CO2 (95:5, vol/vol) for cDMEM containing 2 × 105 cells and maintained for 21
4 h 37°C. Cells were passaged by trypsinization at 70% d for complete differentiation of Caco-2 cells. The day
to 90% confluence (6 ± 1 d post-seeding), and passages of the experiment, Caco-2 cells were washed with sterile
between 35 and 37 were used in this experiment. Medium DPBS, and 0.5 mL of the kefir and kefir-like fermented
was replaced every 48 h except for a serum- and antibi- products were added to the apical side of the insert to
otic-free medium 1 d before all experiments. All reagents a final concentration of 10% kefir. Measurements were
were purchased from Gibco (Thermo Fisher Scientific). taken at 0 h and 8 h, expressed as the ratio of TEER at 8
h compared with 0 h, and then normalized to the negative
Cytotoxicity of Kefir-Like Fermented Product control (DPBS) value.
on Caco-2 Cells
qPCR Analysis of Tight Junction Proteins on Caco-2
Cytotoxicities of kefir and kefir-like products were Cells of Kefir and Kefir-Like Fermented Products
measured to evaluate their potential effects in intestinal
cells. For this measurement, we used fully differentiated In addition to TEER, gene expression of tight junction
Caco-2 cells, as a model of intestinal epithelium, seeded proteins was assessed after the 8-h treatment with the
in 24-well plates (Corning). Cytotoxicity was determined kefir and kefir-like fermented products on Caco-2 cells.
by the lactate dehydrogenase (LDH) assay (Cytotoxicity RNA was extracted from Caco-2 cells after treatment with
Detection Kit LDH, Roche Diagnostics) following the Trizol (Sigma Aldrich) and samples were stored at −80°C
manufacturer’s procedure. Briefly, at 4 h and 8 h, 100 µL until RNA extraction following the manufacturer’s pro-
of cell medium from each well was transferred to a sterile tocol. The Bio-Rad (Hercules, CA) iScript Reverse Tran-
96-well plate. Dulbecco’s PBS (DPBS) was used as the scription Kit was used for reverse transcription, where
negative control, and 1% Triton X-100 was used as the the cDNA was diluted 1:10 after quality analysis and
positive control. The plate was centrifuged at 25°C for 10 quantified with a μDrop plate (Thermo Fisher Scientific).
min at 143 × g. The LDH reagent was prepared by mixing cDNA was amplified using the SsoAdvanced Universal
60 µL of LDH catalyst, 2.7 mL of dye solution, and 2.76 SYBR Green Supermix (Bio-Rad). Genes were amplified
mL of DPBS. In a new 96-well plate, 10 µL of the super- using PrimerPCR SYBR Green Assays (Bio-Rad) and
natant, 40 µL of DPBS buffer, and 50 µL of LDH reagent primed with human primers for claudin-1 (Cldn-1), oc-
were added to each well. The plate was incubated at 25°C cludin (Ocl-1), and zonula occludens (Zo-1) normalized
for 30 min in the dark. Absorbance was recorded at 490 to actin in a CFX96 Touch System (Bio-Rad). Analysis
nm. Samples were done in triplicate. To determine the was performed using CFX Maestro Software (Bio-Rad),
percent of cytotoxicity, the following equation was used: where specificity and efficiency were assessed from melt
curves.

% Cytotoxicity =
(Abs sample
− Abs negative ) ×100, Statistical Analysis
(Abs positive
− Abs negative ) Statistical analysis was done in GraphPad Prism V9.1
(San Diego, CA). Metagenomics are single analysis,
where Abssample was the absorbance of the kefir, Absnegative PCR has n = 5, and all other values are reported as trip-
was absorbance of PBS, and Abspositive was the absorbance licates with mean ± SD. Analysis of variance was done
of 1%Triton. Three independent trials were done for each where assumptions were met and followed with post-hoc
kefir-like fermented product. Tukey for differences between groups. Live cell counts
were log-transformed and analyzed for significant dif-
TEER of Kefir and Kefir-Like Fermented Products ference in linear regression slopes between W0 and W4.
Standard curves were used for all spectrophotometric
To determine the potential effect of the kefir and assays except DPPH. Due to the curved nature of the
kefir-like fermented products on the status of the epi- DPPH standard curve and range of radical scavenging
thelial membrane barrier function, we determined tran- percentage, an ANOVA was done to compare Trolox per-
sepithelial electrical resistance (TEER) as described by cent radical scavenging activity to the kefir supernatants.
Carullo et al. (2020). The TEER was measured with the A 95% significance value was use for all statistics, and
EVOM Epithelial Voltohmmeter with an STX2 probe P-values are reported. Asterisks on results figures for P-
Journal of Dairy Science Vol. 107 No. 7, 2024
 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4264
values are as follows: **** P < 0.0001, *** P < 0.001,
** P < 0.01, and * P < 0.05.

supernatant
4.03 ± 0.01
4.11 ± 0.01
3.92 ± 0.01
4.18 ± 0.01
3.95 ± 0.01
3.95 ± 0.01
pH of
RESULTS

Fermentation Conditions and Kefir Grain Growth

pH of kefir/kefir-like
Compositional analysis of fermentation media is im-

3.72 ± 0.02
3.78 ± 0.02
3.50 ± 0.02
3.64 ± 0.03
3.90 ± 0.03
3.90 ± 0.02
portant when analyzing viability of starter cultures and

product
probiotics, especially in their application for production
of fermented dairy beverages. The by-products are pri-
marily compromised of water with a variety of protein,
carbohydrates (lactose), and fat, as seen in Table 1. We
found that SM, WM, and BM had the highest total pro-

pH of fermentation
tein, and WM and BM had the highest fat content.

6.69 ± 0.01
6.55 ± 0.01
6.54 ± 0.02
4.32 ± 0.01
5.64 ± 0.01
6.65 ± 0.01
Kefir grain weight was tracked at every media change

media
(Figure 1). Those weights were used as indicators of
grain growth under medium conditions. Changes in grain
weight (at D14) are related to kefir grain samples taken
from the grain. The WM and SM grains achieved the
highest biomass grain production. Variability in a grain

0.47 ± 0.01
0.48 ± 0.00
0.44 ± 0.00
0.36 ± 0.01
0.18 ± 0.01
0.13 ± 0.01
Ash (%)
weight can be caused by medium being captured in the
grain matrix as seen in the WM grain D21 and D23.
Growth rates were calculated for the kefir grains as an
indicator of grain viability over the 4-wk period (Figure

Carbohydrates
1). The grains grown in by-products did not exhibit the

3.45 ± 0.07
5.19 ± 0.08

4.14 ± 0.03
3.59 ± 0.13
3.76 ± 0.01
4.03 ± 0.11
same growth rates as those from WM and SM grains. The

(%)
WM grain grew 0.904 ± 0.21 g/48-h fermentation, fol-
lowed by the SM grain (0.713 ± 0.19 g), the BM grain
4.13 ± 0.05
0.14 ± 0.01
1.01 ± 0.12
0.09 ± 0.01
0.10 ± 0.01
0.14 ± 0.01
Fat (%)
Table 1. Macromolecule compositional analysis of fermentation substrates1

Protein (%)
3.91 ± 0.03
6.63 ± 0.23
3.72 ± 0.02
0.31 ± 0.00
0.29 ± 0.01
0.26 ± 0.01
11.97 ± 0.10
9.22 ± 0.10

4.90 ± 0.03
4.15 ± 0.12
3.62 ± 0.02
9.21 ± 0.3
TS (%)

Values represent the mean ± SD of 3 replicates.

Moisture (%)
88.03 ± 0.10
90.78 ± 0.10
90.78 ± 0.29
95.10 ± 0.03
95.84 ± 0.12
95.71 ± 0.02
SWP+ MFGM

Figure 1. Grain growth weight over 4-wk fermentation. Dips ob-

AWP

served at d 15 are reflective of a sample of kefir grain taken from the

SWP
Item
WM

BM
SM

whole grain for additional analysis.

1

Journal of Dairy Science Vol. 107 No. 7, 2024

 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4265

(0.355 ± 0.26 g), the AWP grain (0.161 ± 0.13 g), the
SWP grain (0.216 ± 0.10 g), and the SWP+MFGM grains
(0.148 ± 0.02 g).

Kefir Grain Microbiota with Live Cell Counts,

Shotgun Metagenomics, and SEM

Live cell counts of the kefir and kefir-like products

were determined across the 4-wk fermentation time
(Figure 2). On lactobacilli-selective medium, consistent
growth occurred in all the fermented products, ranging
from 6.5 to 9 log cfu/mL. The BM kefir-like product had
a significant increase (slope [m] = 0.3050, P = 0.0003)
in lactobacilli counts from W0 to W4. The SWP+MFGM
product also had a significant increase (m = 0.1900, P =
0.0102) from W0 to W4. On M17 medium, AWP had no
countable lactococci, whereas BM kefir-like product had
a significant increase (m = 0.3550, P < 0.0001) in live
cell counts, followed by SWP (m = 0.2142, P = 0.0056);
WM had a marginally significant decrease (m = −0.2158,
P = 0.076) in lactococci from W0 to W4. In yeast-selec-
tive medium, WM kefir (m = −0.3625, P < 0.0001) and
SWP kefir-like product had significant decreases (m =
−0.1792, P = 0.0002) in live cell counts at W4 from W0.
The results of the shotgun metagenomic analysis are
shown in Figure 3. The threshold for plotting a species
was a relative abundance of 0.5%. The microbial diver-
sity of the kefir grains underwent significant changes
during the 4-wk fermentation, regardless of the growth
media (Figure 4). At the family level, Lactobacilla-
ceae was the predominant grouping, making up 99.7%
for WM W0, 99.8% for WM W4, 99.8% for SM W0,
99.6% for SM W4, 99.8% for BM W0, 99.7% for BM
W4, 99.8% for AWP W0, 99.7% for AWP W4, 99.7% for
SWP W0, 98.8% for SWP W4, 99.1% for SWP+MFGM
W0, and 98.8% for SWP+MFGM W4. Lactobacillus
kefiranofaciens was detected as the most abundant spe-
cies across all 12 kefir grains, with the following relative
abundances: 94% (AW W0), 70% (AW W4), 71% (BM
W0), 69% (BM W4), 66% (SM W0), 57% (SM W4),
74% (SWP W0), 64% (SWP W4), 73% (SWP+MFGM
W0), 69% (SWP+MFGM W4), 68% (WM W0), and 65%
(WM W4). Lactobacillus helveticus, Lentilactobacillus
kefiri, and Lactoplantibacillus paraplantarum were also
identified with high relative abundances in all samples.
The only lactococcal culture identified metagenomically
was Oenococcus oeni, which was detected in 11 grain
samples (all but AW W0) with a relative abundance
above 0.5%. Eukaryotes were only quantifiable in AW
W4, SWP W0, SWP W4, and SWP+MFGM W0. Con-
tigs were identified as Candida spp. (AWP W4) and
Phaeodactylum tricornutum (SWP W0), Spathaspora Figure 2. Live cell counts of kefir and kefir-like products over a 4-wk
passalidarum (SWP+MFGM W0), and other unspecified fermentation for (A) lactobacilli, (B) lactococci, and (C) eukaryotes.
Error bars represent SD.
eukaryotic species (SWP W4).
Journal of Dairy Science Vol. 107 No. 7, 2024
 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4266

Figure 3. Relative abundance of species detected in the kefir grain by shotgun metagenomics. Grains grown in (A) whole milk, (B) skim milk,
(C) buttermilk, (D) acid whey permeate, (E) sweet whey permeate, and (F) sweet whey permeate +1.5% MFGM. Letters followed by the number 1
are at the start of the fermentation, wk 0. Letters followed by the number 2 are from wk-4 fermented grains. Threshold for plotting for 0.5% relative
abundance.

Scanning electron microscope (SEM) images (Figure pockets within the grain surface matrix). The SWP grain
5) of kefir grain surfaces depict yeast, bacilli, cocci, and (Figure 5E) depicts both budding yeast and bacilli. Final-
extracellular material presumed to be EPS layers on the ly, in the SWP+MFGM grain (Figure 5F), a lactococcal
surface of the kefir grains. The WM grain (Figure 5A) cluster can been seen in the center of the image, and this
exhibited a dense composition with abundant EPS lay- was the only sample where Lactococcus microorganisms
ers. The SM grain (Figure 5B) image depicts chains of were seen on the grain surface.
bacilli and extracellular EPS. The BM grain (Figure 5C)
displayed a flatter surface solely containing bacilli and Kefir Grain Surface Adhesiveness and EPS
limited extracellular EPS. The AW grain (Figure 5D) Production from the Grain
showed a significant amount of yeast with more dimen-
sionality in the surface, exhibited in the surface charging Figure 6 shows the results obtained from the adhesive-
(difference in greyscale in the image depicts peaks and ness evaluation on the grains fermented with the different

Journal of Dairy Science Vol. 107 No. 7, 2024

 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4267

Figure 3 (Continued). Relative abundance of species detected in the kefir grain by shotgun metagenomics. Grains grown in (A) whole milk, (B)
skim milk, (C) buttermilk, (D) acid whey permeate, (E) sweet whey permeate, and (F) sweet whey permeate +1.5% MFGM. Letters followed by the
number 1 are at the start of the fermentation, wk 0. Letters followed by the number 2 are from wk-4 fermented grains. Threshold for plotting for
0.5% relative abundance.

by-products after 4 wk. Although the most adhesive grain Table 2 highlights the estimated EPS yield and the
surfaces were grown in SWP and SWP+MFGM, no sig- quantitative colorimetric results of the reducing and non-
nificant differences were detectable between these treat- reducing sugars from 5 g of grain grown in the different
ments. Likewise, the more traditional milk kefirs (WM media. The SM grain had the highest sugar content of 19
and SM) showed no significant difference in surface µg per 1 mg/mL of hydrated EPS. The kefir grain grown
adhesiveness properties. The only positive adhesiveness in SWP+MFGM had the most EPS extracted, with an
value occurred for BM, indicating that EPS production estimated 6.6 µg of estimated non-reducing sugar from 1
from the grain may be limited. mg/mL of hydrated EPS extract.

Journal of Dairy Science Vol. 107 No. 7, 2024

 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4268

Figure 4. Relative abundance of kefir grain species most commonly identified across grain metagenomics, grown in varying fermentation sub-
strates, at weeks 0 (W0) and 4 (W4) of fermentation.

Antibacterial Activity of Kefir and Kefir- and kefir-like supernatants, there was at least a 5.7 log
Like Supernatants cfu/mL reduction of E. coli across all treatments. Skim
milk kefir supernatant had the lowest viable E. coli cells,
Results of the antibacterial activity against surrogate observing ~6.92 log reduction in 24 h. For L. innocua,
microorganisms shown in the live cell count graphs the initial inoculum culture was estimated to be 9 ± 0.11
(Figure 7A and B) demonstrate the difference in log log cfu/mL (Figure 7B). All 6 kefir supernatants and
colony-forming units from starter surrogate cultures kefir-like supernatants were able to achieve a 5.8 log cfu/
upon treatment with kefir and kefir-like supernatants. At mL reduction for both L. innocua, with SM kefir super-
24 h, the inoculation concentration of E. coli was ~9.22 natant having the largest log reduction of 6.68 log cfu/
± 0.04 log cfu/mL (Figure 7A). When treated with kefir mL. Generally, only 2 to 3 log cfu/mL of each surrogate
Journal of Dairy Science Vol. 107 No. 7, 2024
 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4269

Figure 5. Scanning electron microscopy images from kefir grain surfaces grown in (A) whole milk, (B) skim milk, (C) buttermilk, (D) acid whey
permeate, (E) sweet whey permeate, and (F) sweet whey permeate with 1.5% MFGM.

managed to survive when exposed to kefir supernatants product supernatant (Figure 8A). Whole milk kefir su-
and kefir-like supernatants. This finding suggests a lack pernatant was not significantly different than the Trolox
of significant difference in the antibacterial activity be- standard. Radical scavenging activities of SM kefir and
tween the fermented products. BM kefir-like supernatant were greater than the Trolox
standard. The SWP and SWP+MFGM kefir-like super-
Antioxidant Activity of Kefir and Kefir- natants were not significantly different from each other.
Like Supernatants For ABTS, the relative TEAC for each kefir and kefir-
like supernatant is plotted on Figure 8B with pair-wise
The antioxidant activity results reflected the antioxi- comparisons. The SWP kefir-like supernatant had the
dant potential of the kefir-like products compared with lowest TEAC (14.3 ± 1.4 µM) and was significantly dif-
those found in traditional kefirs (WM and SM). The ferent than that of the SWP+MFGM kefir-like superna-
radical scavenging activity (DPPH method) of the kefir tant (168.5 ± 12.3 µM). Skim milk had the highest TEAC
and kefir-like products was assessed by comparing to a (365.3 ± 5.1 µM). We found no significant difference
Trolox standard (400 µM) and to each kefir and kefir-like between AWP and BM kefir-like supernatants.

Journal of Dairy Science Vol. 107 No. 7, 2024

 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4270
Table 2. Yield and estimated EPS from 5 g of kefir grain and a 1 mg/mL EPS sample from each grain1

Estimated
EPS yield from 5 g Phenol sulfuric total DNS2 reducing nonreducing
Kefir or kefir-like supernatant of grain (mg) phenolics (µg) sugar (µg) sugars (EPS; µg)
WM 367.6 ± 2.03 14.4 ± 0.31 10.6 ± 0.40 3.79 ± 0.71
SM 380.9 ± 1.87 32.3 ± 0.90 12.9 ± 1.5 19.4 ± 1.7
BM 352.0 ± 1.72 24.1 ± 1.2 9.7 ± 0.38 14.4 ± 0.96
AWP 360.5 ± 2.03 14.4 ± 0.24 11.0 ± 0.18 3.45 ± 0.25
SWP 436.6 ± 1.72 16.0 ± 3.5 11.1 ± 2.4 4.94 ± 5.3
SWP+MFGM2 453.5 ± 2.14 13.6 ± 0.61 7.0 ± 0.77 6.66 ± 1.07
1
Means are taken from triplicates and reported ±SD.
2
DNS = dinitrosalicylic acid.

Kefir and Kefir-Like Products’ Impact BM kefir-like products and all other kefir and kefir-like
on Caco-2 Cytotoxicity products (P < 0.0001 for all). Other significant differ-
ences were between SM kefir and AWP kefir-like prod-
Cytotoxicity (Figure 9) showed a trending decrease in uct (P = 0.0303), between SM kefir and SWP kefir-like
effect between 4 and 8 h on Caco-2 cells. The range of per- products (P = 0.0017), between AWP and SWP+MFGM
cent cytotoxicity of kefir and kefir-like products ranged kefir-like products (P = 0.0004), and between SWP and
from 0.81% (BM, 8 h) to 28% (WM, 4 h). Compared SWP+MFGM (P < 0.0001). There was no significant
with 4 and 8 h, only WM kefir (10%) and BM kefir-like difference at 4 h between SM kefir and SWP+MFGM
product (10%) were significantly different. At 4 h the kefir-like product (P = 0.1442) or between AWP and
mean cytotoxicities were as follows: 26.70% for WM, SWP kefir-like products (P = 0.5213). At 8 h, the mean
16.90% for SM, 0.8% for BM, 13.74% for AWP, 12.22% cytotoxicities were as follows: 15.44% for WM, 17.31%
for SWP, and 18.23% for SWP+MFGM. There was sig- for SM, 0% for BM, 11.65% for AWP, 11.72% for SWP,
nificant difference between WM kefir and all the other and 17.28% for SWP+MFGM. There were significant
kefir and kefir-like products (all P < 0.0001), as well as differences between BM and all other kefir and kefir-
like products (P < 0.0001), WM kefir and AWP kefir-
like product (P = 0.0024), WM and SWP (P = 0.0028),
SM kefir compared with both AWP and SWP kefir-like
products (P < 0.0001), SWP and SWP+MFGM kefir-like
products (P < 0.0001), and SWP kefir-like product and
SWP+MFGM kefir-like products (P < 0.0001). There
were no significant differences between WM and SM ke-
firs (P = 0.1803), WM kefir and SWP+MFGM kefir-like
product (P = 0.1949), SM kefir and SWP+MFGM kefir-
like product (P > 0.9999), and AWP and SWP kefir-like
products (P > 0.999).
A Caco-2 cell model was used to assess the membrane
barrier capability of intestinal cells when exposed to kefir
and kefir-like fermented products, with TEER measured
at the start of treatment and at the end of the 8-h exposure
to kefir. The TEER values were normalized to the PBS-
negative and expressed as the ratio of change in TEER.
Ratios greater than 1 indicate increased TEER values, as
observed in SM kefir and AWP, SWP, and SWP+MFGM
kefir-like products (Figure 10). These ratios were signifi-
cantly higher (P < 0.05) than those of the positive control
(1% Triton). Among these, SWP+MFGM kefir-like prod-
uct and SM kefir had the greatest positive impact on the
Caco-2 cell line barrier, with TEER values higher than
the negative control (PBS).
To determine whether the observed increase in the
Figure 6. Kefir grain surface adhesiveness analyzed with ANOVA
and Tukey pairwise comparisons. Asterisks indicate significant differ- TEER values was associated with an induction in the
ence: **** P < 0.0001. Error bars represent SD. gene expression of tight junction proteins, we assessed
Journal of Dairy Science Vol. 107 No. 7, 2024
 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4271

Figure 7. Antibacterial activity of kefir and kefir-like products reported as log reduction against Escherichia coli ATCC 25922 (A) and Listeria
innocua ATCC 25922 (B). Data represent the optical density spectra with kefir and kefir-like products inoculated with surrogate pathogens and the
plate counts in log cfu/mL. Error bars represent SD.

gene expression of Cldn-1, Ocln-1, and Zo-1 and com- population (Mazzetto et al., 2022). Finding new uses for
pared with the negative (PBS) control (Figure 11). For these by-products could help reduce the environmental
Cldn1, there was a significant (P < 0.0001) induction of burden.
the gene expression in the SWP and SWP+MFGM kefir- The primary negative trait in whey-based fermented
like products. Ocln-1 gene expression was significantly beverages is texture, as they are primarily water-based,
(P < 0.0001) upregulated in SWP+MFGM kefir-like which can be solved through the addition of thickening
product. Conversely, WM and SM kefir, and SWP and agents (de Matos Reis et al., 2021). Exopolysaccha-
SWP+MFGM kefir-like products had a significant (P < rides have an application as a functional ingredient as
0.0001) increase in Zo-1 expression. Although not sig- a thickening agent. In stirred yogurt, Rawson and Mar-
nificant, WM, SWP, and SWP+MFGM showed signs of shall (1997) found that increased “ropy” strains of lactic
inducing Zo-1 expression. acid bacteria (LAB; i.e., EPS-producing bacteria) in the
yogurt culture was associated with a decrease adhesive-
DISCUSSION ness of the product; the authors hypothesized that this
was related to the protein matrix. Because of this, the
The annual waste generated by the dairy industry is difference in adhesiveness on the surface of kefir grains
estimated to be ~25 billion pounds, with most of that may be associated with the survival of EPS-producing
waste occurring during the separation of milk into other microorganisms, the primary one from the metagenomic
dairy products (Campbell and Feldpausch, 2022). This results in this research and previous literature being L.
waste not only exerts a negative impact on the environ- kefiranofaciens. Besides the TPA adhesiveness implying
ment but also casts a shadow over the dairy industry, viability of EPS-producing bacteria, EPS themselves can
drawing concerns from consumers and the general positively affect rheological and sensory properties of the

Journal of Dairy Science Vol. 107 No. 7, 2024

 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4272

Figure 8. Antioxidant activity measured with (A) 2,2-diphenyl-1-picrylhydrazyl assay and (B) 2,2′-azino-bis (3-ethylbenzothiazoline-6-sulfonic
acid) assays, analyzed with ANOVA and Tukey pairwise comparisons. Eq. = equivalence. Asterisks indicate significant difference: **** P < 0.0001,
** P < 0.01, and * P < 0.05. Error bars represent SD.

final product and are being studied for their independent EPS biofilms. The addition of milk fat can be seen as
functional properties (Hassan et al., 2003). an additional protective layer for the kefir grain bacteria
The SEM images of kefir grain surfaces depict cocci, and may explain why the products fermented with BM
bacilli, and yeast interactions. Cellular matter present and SWP+MFGM had higher live cell count viability
in the SEM images can be classified as potential milk throughout the fermentation and were able to increase
fat or phospholipids protecting the grain microbiota and biomass (grain growth) during the fermentation. The
difference in biomass and grain weight captured in this
research suggests that the kefir grain matrix may cap-
ture proteins within the EPS barrier, as evidenced by
the difference in biomass and grain weight results. Al-
though the deproteinized by-products have comparable
viability, indicative of microbial viability, they displayed
lower biomass weight. The difference in biomass and
grain weights captured in this research can lead us to
hypothesize that the kefir grain matrix captures proteins
within the EPS barrier. Important metabolites of some
kefir grain microorganisms, EPS account for up to half of
the structure of the kefir grain matrix (La Riviére et al.,
1967) and may play a role in the protection and survival
of kefir grain microorganisms.
Metagenomic data were used to identify specific cul-
tures in the kefir grain. However, identification in the
kefir grain does not correlate to survival in the grain
and presence in the kefir product, as seen in González-
Figure 9. Cytotoxicity from the lactate dehydrogenase assay of kefir Orozco et al. (2023). Likewise, metagenomic data only
and kefir-like products at 4 and 8 h. Results were analyzed with a paired detects DNA and does not distinguish between living
t-test between 4 and 8 h per each medium. Asterisks indicate significant
difference: **** P < 0.0001, *** P < 0.001. Error bars represent SD. and dead cultures, which may explain different cultures

Journal of Dairy Science Vol. 107 No. 7, 2024

 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4273

identified at W4 in the data that were not present in W0.

Lactobacillus kefiranofaciens was the most abundant
bacteria species across all samples and the species most
attributed to traditional kefir, although not present in
commercially produced kefir (García Fontán et al., 2006;
Biçer et al., 2021). Selective media in the live cell counts
can allude to a more diverse microbial population than
just metagenomics data alone. On lactococcal selective
medium, AWP had no countable colonies. Research with
Lactococcus lactis indicates that when the pH of an en-
vironment is below 5, cell viability is decreased, along
with other environmental factors (Molina-Gutierrez et
al., 2002). The low pH of acid whey could cause stress-
induced cell deaths. When cocci species were identified
metagenomically and confirmed at the family level on
M17 medium, Oenococcus oeni was isolated. Oenococ-
cus oeni is commonly found in water kefir (Verce et al.,
2019) and is used in wine fermentation (Mills et al., 2005).
The species undergoes malolactic fermentation, usually
caused by the presence of malic acid sourced from a
cow’s diet (Wang et al., 2009). The presence of traces of
malic acid in processed milk that is then fermented with
kefir microbiota is not well understood; further research
with this species in fermented milk is needed. Eukaryotes
are an important part of kefir grain microbiota and can
contribute to the alcohol content in kefir, and kefir yeast
has been used for whey-based alcohol production (Atha-

Figure 10. Transepithelial electrical resistance (TEER) values rela-

tive to 0 h and 8 h, compared with the positive control analyzed with Figure 11. Changes in gene expression of tight junction proteins (A)
ANOVA and Tukey pairwise comparisons. Asterisks indicate significant Cldn-1, (B) Ocln-1, and (C) Zo-1 analyzed with ANOVA and Tukey
difference: *** P < 0.001, ** P < 0.01, and * P < 0.05. Error bars rep- pairwise comparisons. Asterisks indicate significant difference: **** P
resent SD. < 0.0001 and * P < 0.05. Error bars represent SD.

Journal of Dairy Science Vol. 107 No. 7, 2024

 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4274
nasiadis et al., 2002, 2004). Spathaspora passalidarum cells in a time and dose response and has a pro-apoptotic
and Candida spp. were identified from the kefir grain in effect on those 2 cell lines. In this study, kefir and kefir-
this research. Due to library accessibility, it is very pos- like products (except for buttermilk kefir-like product)
sible that what were presumably identified as Candida demonstrated antiproliferation and pro-apoptotic effects
spp. in this study could be Kluyveromyces spp., which on Caco-2 cells.
are more commonly isolated from kefir grains (Diosma The results observed in the metagenomic analysis and
et al., 2014). The diverse range of microorganisms able live cell count viability, supplemented with the bioactive
to survive in the various by-products is applicable for results, highlight potential applications for the waste
fermented dairy product starter cultures. streams. Consumers willing to try kefir and kefir-like
Functional health properties associated with kefir and products are likely to be interested in the associated
kefir-like products are mainly attributed to the kefir mi- health benefits, highlighting the importance of screening
croorganisms and the wide range of metabolites produced studies as seen in this research. More research regarding
during the fermentation (Bengoa et al., 2019), including the bioactivities of kefir and kefir-like products should
EPS. Antimicrobial activity observed in kefir can be at- be done in in vitro and in vivo models.
tributed to antimicrobial substances such as EPS from
L. kefiranofaciens (Jeong et al., 2017; Abdalla et al., CONCLUSIONS
2021), bacteriocins from L. helveticus, and organic acids
from Lactobacillus plantarum (Hassan et al., 2020), all By-products from dairy processing are a large source
of which were identified in the kefir grains used in this of waste and have a negative environmental impact, caus-
study. In a study by Al-Mohammadi et al. (2021), kefir ing consumers to turn to alternative dairy products. We
was able to reduce Staphylococcus aureus and Salmo- propose that finding applications for by-products reduces
nella Typhimurium by 5 log cfu. Similarly, antioxidant waste and improves sustainability while fostering in-
activity from kefir can be attributed to bioactive peptides novation for novel products. Microorganisms identified
released from the fermentation with microorganisms. from kefir grain bacteria should be considered for fer-
High antioxidant activity in fermented dairy products can mentation substrates for novel probiotic dairy beverages
be attributed to the microorganisms breaking down whey or other fermented dairy products. Commercially used
peptide chains, potentially releasing cysteine, which is strains of LAB for fermented dairy foods are limited and
then able to donate hydrogen atoms to DPPH radicals, typically consist of bacilli such as Lactobacillus bulgari-
resulting in spectrophotometric changes observed in the cus, Lactobacillus helveticus, and Bifidobacterium spp.,
DPPH assay, as previously observed with Lactobacillus and some cocci, including Streptococcus thermophilus
acidophilus (Sabokbar and Khodaiyan, 2016). Although and Lactococcus lactis (Biçer et al., 2021). The limita-
L. acidophilus was not identified in the kefir grain tions of commercially used LAB strains provide a unique
metagenomic analysis in this research, there is potential opportunity for kefir grains—a complex microbial con-
that other LAB species undergo the same metabolic path- sortia—to valorize dairy processing by-products through
way in deproteinated media (whey permeates and but- fermentation. Further research involving the neutraliza-
termilk). The complex microbiota of kefir grains and the tion of acid whey and the microbial survival of kefir
different metabolic pathways of both yeast and bacteria grain microorganisms to reduce an additional by-product
can provide fermented dairy foods with bioactive mol- through fermentation, as well as product development
ecules that can be used for antimicrobial or antioxidant with sweet whey permeate for a novel fermented probi-
functions. otic beverage, should be performed.
Intestinal epithelial cells provide a protective barrier
from the exterior environment, preventing harmful sub- NOTES
stances, pathogens, and toxins from infiltrating the intes-
tinal lumen. Tight junction proteins aid in the protection The authors thank the J. T. “Stubby” Endowed Chair
of the epithelial membrane barrier primarily through in Dairy Foods at The Ohio State University (Columbus,
intercellular communications and by regulating paracel- OH) for funding this research and Molly J. Davis from the
lular transport (Bhat et al., 2019). Gut damage can be department of Food Science and Technology at The Ohio
caused by foodborne illnesses, irritable bowel disease State University (Columbus, OH) for editing the manu-
and other inflammatory bowel diseases such as Crohn’s script. Supplemental material for this article is available
disease or ulcerative colitis, as well as aging and stress at https:​/​/​doi​.org/​10​.6084/​m9​.figshare​.25733517​.v1. Be-
(Anderson et al., 2010). Based on the results achieved cause no human or animal subjects were used, this analy-
in this research, SM kefir and kefir from SWP+MFGM sis did not require approval by an Institutional Animal
can potentially sustain a tight epithelial layer. Khoury et Care and Use Committee or Institutional Review Board.
al. (2014) suggested that kefir affects Caco-2 and HT-29 The authors have not stated any conflicts of interest.
Journal of Dairy Science Vol. 107 No. 7, 2024
 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4275

Abbreviations used: ABTS = 2,2′-azino-bis (3-ethyl- Bengoa, A. A., C. Iraporda, G. Garrote, and A. Abraham. 2019. Kefir
micro-organisms: Their role in grain assembly and health properties
benzothiazoline-6-sulfonic acid); AWP = acid whey per- of fermented milk. J. Appl. Microbiol. 126:686–700. https:​/​/​doi​.org/​
meate kefir-like supernatant; BM = buttermilk kefir-like 10​.1111/​jam​.14107.
supernatant; cDMEM = complete Dulbecco’s modified Bhat, A. A., S. Uppada, I. W. Achkar, S. Hashem, S. K. Yadav, M. Shan-
mugakonar, H. A. Al-Naemi, M. Haris, and S. Uddin. 2019. Tight
Eagle’s medium; D7 = d 7; DNS = dinitrosalicylic acid; junction proteins and signaling pathways in cancer and inflamma-
DPBS = Dulbecco’s PBS; DPPH = 2,2-diphenyl-1-pic- tion: A functional crosstalk. Front. Physiol. 9:1942. https:​/​/​doi​.org/​
rylhydrazyl; EPS = exopolysaccharide; gDNA = genomic 10​.3389/​fphys​.2018​.01942.
Biçer, Y., A. Ezgi Telli, G. Sönmez, G. Turkal, N. Telli, and U. Gürkan.
DNA; LAB = lactic acid bacteria; LDH = lactate dehy- 2021. Comparison of commercial and traditional kefir microbiota
drogenase; m = regression line slope; MFGM = milk fat using metagenomic analysis. Int. J. Dairy Technol. Wiley Online
globule membrane; MPL = milk phospholipids; MRS = Library. Accessed Aug. 9, 2023. https:​/​/​onlinelibrary​.wiley​.com/​doi/​
full/​10​.1111/​1471​-0307​.12789.
De Man, Rogosa, Sharpe; OD = optical density; PPB = Campbell, C. G., and G. L. Feldpausch. 2022. Invited review: The con-
potassium phosphate buffer; RBCA = Rose Bengal chlor- sumer and dairy food waste: An individual plus policy, systems, and
amphenicol agar; SEM = scanning electron microscope; environmental perspective. J. Dairy Sci. 105:3736–3745. https:​/​/​doi​
.org/​10​.3168/​jds​.2021​-20994.
SM = skim milk kefir supernatant; SWP = sweet whey Carullo, G., P. Governa, U. G. Spizzirri, M. Biagi, F. Sciubba, G.
permeate kefir-like supernatant; SWP+MFGM = sweet Giorgi, M. R. Loizzo, M. E. Di Cocco, F. Aiello, and D. Restuc-
whey permeate with milk fat globule membrane kefir-like cia. 2020. Sangiovese cv pomace seeds extract-fortified kefir exerts
anti-inflammatory activity in an in vitro model of intestinal epithe-
supernatant; TEER = transepithelial electrical resistance; lium using Caco-2 cells. Antioxidants 9:54. https:​/​/​doi​.org/​10​.3390/​
TPA = texture profile analysis; W0–W4 = wk 0 to wk 4 of antiox9010054.
fermentation; WM = whole milk kefir supernatant. Clausen, P. T. L. C., F. M. Aarestrup, and O. Lund. 2018. Rapid and
precise alignment of raw reads against redundant databases with
KMA. BMC Bioinformatics 19:307. https:​/​/​doi​.org/​10​.1186/​s12859​
REFERENCES -018​-2336​-6.
de Matos Reis, S., G. R. L. Mendes, B. M. A. C. Mesquita, W. J. N.
Abdalla, A. K., M. M. Ayyash, A. N. Olaimat, T. M. Osaili, A. A. Al- Lima, C. A. F. D. Pinheiro, F. A. O. Ruas, G. L. M. Santos, and I.
Nabulsi, N. P. Shah, and R. Holley. 2021. Exopolysaccharides as V. Brandi. 2021. Development of milk drink with whey fermented
antimicrobial agents: Mechanism and spectrum of activity. Front. and acceptability by children and adolescents. J. Food Sci. Technol.
Microbiol. 12:664395. https:​/​/​doi​.org/​10​.3389/​fmicb​.2021​.664395. 58:2847–2852. https:​/​/​doi​.org/​10​.1007/​s13197​-021​-05003​-w.
Al-Mohammadi, A.-R., R. A. Ibrahim, A. H. Moustafa, A. A. Ismaiel, A. Diosma, G., D. E. Romanin, M. F. Rey-Burusco, A. Londero, and G.
Abou Zeid, and G. Enan. 2021. Chemical constitution and antimicro- L. Garrote. 2014. Yeasts from kefir grains: Isolation, identification,
bial activity of kefir fermented beverage. Molecules 26:2635. https:​ and probiotic characterization. World J. Microbiol. Biotechnol.
/​/d​ oi​.org/​10​.3390/​molecules26092635. 30:43–53. https:​/​/​doi​.org/​10​.1007/​s11274​-013​-1419​-9.
Anderson, R. C., A. L. Cookson, W. C. McNabb, Z. Park, M. J. Mc- DuBois, M., K. A. Gilles, J. K. Hamilton, P. A. Rebers, and F. Smith.
Cann, W. J. Kelly, and N. C. Roy. 2010. Lactobacillus plantarum 1956. Colorimetric method for determination of sugars and re-
MB452 enhances the function of the intestinal barrier by increasing lated substances. Anal. Chem. 28:350–356. https:​/​/​doi​.org/​10​.1021/​
the expression levels of genes involved in tight junction formation. ac60111a017.
BMC Microbiol. 10:316. https:​/​/​doi​.org/​10​.1186/​1471​-2180​-10​-316. Enikeev, R. 2012. Development of a new method for determination of
Athanasiadis, I., D. Boskou, M. Kanellaki, V. Kiosseoglou, and A. A. exopolysaccharide quantity in fermented milk products and its ap-
Koutinas. 2002. Whey liquid waste of the dairy industry as raw ma- plication in technology of kefir production. Food Chem. 134:2437–
terial for potable alcohol production by kefir granules. J. Agric. Food 2441. https:​/​/​doi​.org/​10​.1016/​j​.foodchem​.2012​.04​.050.
Chem. 50:7231–7234. https:​/​/​doi​.org/​10​.1021/​jf020331q. Erdogan, F. S., S. Ozarslan, Z. B. Guzel-Seydim, and T. Kök Taş. 2019.
Athanasiadis, I., A. Paraskevopoulou, G. Blekas, and V. Kiosseoglou. The effect of kefir produced from natural kefir grains on the intesti-
2004. Development of a novel whey beverage by fermentation nal microbial populations and antioxidant capacities of Balb/c mice.
with kefir granules. Effect of various treatments. Biotechnol. Prog. Food Res. Int. 115:408–413. https:​/​/​doi​.org/​10​.1016/​j​.foodres​.2018​
20:1091–1095. https:​/​/​doi​.org/​10​.1021/​bp0343458. .10​.080.
Avila-Reyes, S. V., C. E. Márquez-Morales, G. R. Moreno-León, A. Gao, J., F. Gu, J. He, J. Xiao, Q. Chen, H. Ruan, and G. He. 2013.
R. Jiménez-Aparicio, M. L. Arenas-Ocampo, J. Solorza-Feria, E. Metagenome analysis of bacterial diversity in Tibetan kefir grains.
García-Armenta, and J. C. Villalobos-Espinosa. 2022. Comparative Eur. Food Res. Technol. 236:549–556. https:​/​/​doi​.org/​10​.1007/​
analysis of fermentation conditions on the increase of biomass and s00217​-013​-1912​-2.
morphology of milk kefir grains. Appl. Sci. (Basel) 12:2459. https:​/​ García Fontán, M. C., S. Martínez, I. Franco, and J. Carballo. 2006. Mi-
/​doi​.org/​10​.3390/​app12052459. crobiological and chemical changes during the manufacture of kefir
Azizkhani, M., P. E. J. Saris, and M. Baniasadi. 2021. An in vitro as- made from cows’ milk, using a commercial starter culture. Int. Dairy
sessment of antifungal and antibacterial activity of cow, camel, ewe, J. 16:762–767. https:​/​/​doi​.org/​10​.1016/​j​.idairyj​.2005​.07​.004.
and goat milk kefir and probiotic yogurt. J. Food Meas. Charact. Gonçalves, C., R. M. Rodriguez-Jasso, N. Gomes, J. A. Teixeira, and I.
15:406–415. https:​/​/​doi​.org/​10​.1007/​s11694​-020​-00645​-4. Belo. 2010. Adaptation of dinitrosalicylic acid method to microti-
Baliyan, S., R. Mukherjee, A. Priyadarshini, A. Vibhuti, A. Gupta, R. P. ter plates. Anal. Methods 2:2046–2048. https:​/​/​doi​.org/​10​.1039/​
Pandey, and C.-M. Chang. 2022. Determination of antioxidants by c0ay00525h.
DPPH radical scavenging activity and quantitative phytochemical González-Orozco, B. D., I. García-Cano, A. Escobar-Zepeda, R.
analysis of Ficus religiosa. Molecules 27:1326. https:​/​/​doi​.org/​10​ Jiménez-Flores, and V. B. Álvarez. 2023. Metagenomic analysis
.3390/​molecules27041326. and antibacterial activity of kefir microorganisms. J. Food Sci.
Bellikci-Koyu, E., B. P. Sarer-Yurekli, Y. Akyon, F. Aydin-Kose, C. 88:2933–2949. https:​/​/​doi​.org/​10​.1111/​1750​-3841​.16614.
Karagozlu, A. G. Ozgen, A. Brinkmann, A. Nitsche, K. Ergunay, Hassan, A. N., R. Ipsen, T. Janzen, and K. B. Qvist. 2003. Microstructure
E. Yilmaz, and Z. Buyuktuncer. 2019. Effects of regular kefir con- and rheology of yogurt made with cultures differing only in their
sumption on gut microbiota in patients with metabolic syndrome: ability to produce exopolysaccharides. J. Dairy Sci. 86:1632–1638.
A parallel-group, randomized, controlled study. Nutrients 11:2089. https:​/​/​doi​.org/​10​.3168/​jds​.S0022​-0302(03)73748​-5.
https:​/​/​doi​.org/​10​.3390/​nu11092089.

Journal of Dairy Science Vol. 107 No. 7, 2024

 McGovern et al.: KEFIR GRAIN VIABILTY IN DAIRY BY-PRODUCTS 4276
Hassan, M. U., H. Nayab, F. Shafique, M. P. Williamson, T. S. Alman- Pereira, T. M. C., L. Z. Côco, A. M. M. Ton, S. S. Meyrelles, M. Campos-
souri, N. Asim, N. Shafi, S. Attacha, M. Khalid, N. Ali, and N. Akbar. Toimil, B. P. Campagnaro, and E. C. Vasquez. 2021. The emerging
2020. Probiotic properties of Lactobacillus helveticus and Lactoba- scenario of the gut–brain axis: The therapeutic actions of the new ac-
cillus plantarum isolated from traditional Pakistani yoghurt. BioMed tor kefir against neurodegenerative diseases. Antioxidants 10:1845.
Res. Int. 2020:8889198. https:​/​/​doi​.org/​10​.1155/​2020/​8889198. https:​/​/​doi​.org/​10​.3390/​antiox10111845.
Hubatsch, I., E. G. E. Ragnarsson, and P. Artursson. 2007. Determination Piermaria, J. A., M. L. de la Canal, and A. G. Abraham. 2008. Gelling
of drug permeability and prediction of drug absorption in Caco-2 properties of kefiran, a food-grade polysaccharide obtained from
monolayers. Nat. Protoc. 2:2111–2119. https:​/​/​doi​.org/​10​.1038/​ kefir grain. Food Hydrocoll. 22:1520–1527. https:​/​/​doi​.org/​10​.1016/​
nprot​.2007​.303. j​.foodhyd​.2007​.10​.005.
Jeong, D., D.-H. Kim, I.-B. Kang, H. Kim, K.-Y. Song, H.-S. Kim, and Rahman, M. M., M. B. Islam, M. Biswas, and A. H. M. Khurshid Alam.
K.-H. Seo. 2017. Characterization and antibacterial activity of a 2015. In vitro antioxidant and free radical scavenging activity of dif-
novel exopolysaccharide produced by Lactobacillus kefiranofaciens ferent parts of Tabebuia pallida growing in Bangladesh. BMC Res.
DN1 isolated from kefir. Food Control 78:436–442. https:​/​/​doi​.org/​ Notes 8:621. https:​/​/​doi​.org/​10​.1186/​s13104​-015​-1618​-6.
10​.1016/​j​.foodcont​.2017​.02​.033. Rawson, H. L., and V. M. Marshall. 1997. Effect of ‘ropy’ strains of
Khoury, N., S. El-Hayek, O. Tarras, M. El-Sabban, M. El-Sibai, and S. Lactobacillus delbrueckii ssp. bulgaricus and Streptococcus ther-
Rizk. 2014. Kefir exhibits anti-proliferative and pro-apoptotic ef- mophilus on rheology of stirred yogurt. Int. J. Food Sci. Technol.
fects on colon adenocarcinoma cells with no significant effects on 32:213–220. https:​/​/​doi​.org/​10​.1046/​j​.1365​-2621​.1997​.00395​.x.
cell migration and invasion. Int. J. Oncol. 45:2117–2127. https:​/​/​doi​ Re, R., N. Pellegrini, A. Proteggente, A. Pannala, M. Yang, and C.
.org/​10​.3892/​ijo​.2014​.2635. Rice-Evans. 1999. Antioxidant activity applying an improved ABTS
Kosmerl, E., D. Rocha-Mendoza, J. Ortega-Anaya, R. Jiménez-Flores, radical cation decolorization assay. Free Radic. Biol. Med. 26:1231–
and I. García-Cano. 2021. Improving human health with milk fat 1237. https:​/​/​doi​.org/​10​.1016/​S0891​-5849(98)00315​-3.
globule membrane, lactic acid bacteria, and bifidobacteria. Micro- Rimada, P. S., and A. G. Abraham. 2001. Polysaccharide production by
organisms 9:341. https:​/​/​doi​.org/​10​.3390/​microorganisms9020341. kefir grains during whey fermentation. J. Dairy Res. 68:653–661.
La Riviére, J. W. M., P. Kooiman, and K. Schmidt. 1967. Kefiran, a novel https:​/​/​doi​.org/​10​.1017/​S0022029901005131.
polysaccharide produced in the kefir grain by Lactobacillus brevis. Rosa, D. D., M. M. S. Dias, Ł. M. Grześkowiak, S. A. Reis, L. L. Con-
Arch. Mikrobiol. 59:269–278. https:​/​/​doi​.org/​10​.1007/​BF00406340. ceição, and M. C. G. Peluzio. 2017. Milk kefir: Nutritional, micro-
Lim, J., M. Kale, D.-H. Kim, H.-S. Kim, J.-W. Chon, K.-H. Seo, H. biological and health benefits. Nutr. Res. Rev. 30:82–96. https:​/​/​doi​
G. Lee, W. Yokoyama, and H. Kim. 2017. Antiobesity effect of .org/​10​.1017/​S0954422416000275.
exopolysaccharides isolated from kefir grains. J. Agric. Food Chem. Ruijssenaars, H. J., F. Stingele, and S. Hartmans. 2000. Biodegradability
65:10011–10019. https:​/​/​doi​.org/​10​.1021/​acs​.jafc​.7b03764. of food-associated extracellular polysaccharides. Curr. Microbiol.
Liu, J.-R., M.-J. Chen, and C.-W. Lin. 2005. Antimutagenic and anti- 40:194–199. https:​/​/​doi​.org/​10​.1007/​s002849910039.
oxidant properties of milk-kefir and soymilk-kefir. J. Agric. Food Sabokbar, N., and F. Khodaiyan. 2016. Total phenolic content and anti-
Chem. 53:2467–2474. https:​/​/​doi​.org/​10​.1021/​jf048934k. oxidant activities of pomegranate juice and whey based novel bever-
Liu, S., S.-Y. Lu, N. Qureshi, H. A. E. Enshasy, and C. D. Skory. 2022. age fermented by kefir grains. J. Food Sci. Technol. 53:739–747.
Antibacterial property and metagenomic analysis of milk kefir. Pro- https:​//​ ​doi​.org/​10​.1007/​s13197​-015​-2029​-3.
biotics Antimicrob. Proteins 14:1170–1183. https:​/​/​doi​.org/​10​.1007/​ Sodini, I., P. Morin, A. Olabi, and R. Jiménez-Flores. 2006. Composi-
s12602​-022​-09976​-8. tional and functional properties of buttermilk: A comparison between
Maeda, H., X. Zhu, S. Suzuki, K. Suzuki, and S. Kitamura. 2004. sweet, sour, and whey buttermilk. J. Dairy Sci. 89:525–536. https:​/​/​
Structural characterization and biological activities of an exopoly- doi​.org/​10​.3168/​jds​.S0022​-0302(06)72115​-4.
saccharide kefiran produced by Lactobacillus kefiranofaciens WT- Tomar, O., G. Akarca, A. Çağlar, M. Beykaya, and V. Gök. 2020. The
2B(T). J. Agric. Food Chem. 52:5533–5538. https:​/​/​doi​.org/​10​.1021/​ effects of kefir grain and starter culture on kefir produced from cow
jf049617g. and buffalo milk during storage periods. Food Sci. Technol. (Campi-
Magalhães, K. T., G. Dragone, G. V. de Melo Pereira, J. M. Oliveira, nas) 40:238–244. https:​/​/​doi​.org/​10​.1590/​fst​.39418.
L. Domingues, J. A. Teixeira, J. B. A. e Silva, and R. F. Schwan. Tung, Y.-T., H.-L. Chen, H.-S. Wu, M.-H. Ho, K.-Y. Chong, and C.-M.
2011. Comparative study of the biochemical changes and volatile Chen. 2018. Kefir peptides prevent hyperlipidemia and obesity in
compound formations during the production of novel whey-based high-fat-diet-induced obese rats via lipid metabolism modulation.
kefir beverages and traditional milk kefir. Food Chem. 126:249–253. Mol. Nutr. Food Res. 62:1700505. https:​/​/​doi​.org/​10​.1002/​mnfr​
https:​/​/​doi​.org/​10​.1016/​j​.foodchem​.2010​.11​.012. .201700505.
Mazzetto, A. M., S. Falconer, and S. Ledgard. 2022. Mapping the carbon Verce, M., L. De Vuyst, and S. Weckx. 2019. Shotgun metagenomics
footprint of milk production from cattle: A systematic review. J. of a water kefir fermentation ecosystem reveals a novel Oenococ-
Dairy Sci. 105:9713–9725. https:​/​/​doi​.org/​10​.3168/​jds​.2022​-22117. cus species. Front. Microbiol. 10:479. https:​/​/​doi​.org/​10​.3389/​fmicb​
Mills, D. A., H. Rawsthorne, C. Parker, D. Tamir, and K. Makarova. .2019​.00479.
2005. Genomic analysis of Oenococcus oeni PSU-1 and its relevance Wang, C., Q. Liu, W. Z. Yang, Q. Dong, X. M. Yang, D. C. He, K. H.
to winemaking. FEMS Microbiol. Rev. 29:465–475. https:​/​/​doi​.org/​ Dong, and Y. X. Huang. 2009. Effects of malic acid on feed intake,
10​.1016/​j​.fmrre​.2005​.04​.011. milk yield, milk components and metabolites in early lactation Hol-
Molina-Gutierrez, A., V. Stippl, A. Delgado, M. G. Gänzle, and R. F. Vo- stein dairy cows. Livest. Sci. 124:182–188. https:​/​/​doi​.org/​10​.1016/​
gel. 2002. In situ determination of the intracellular pH of Lactococ- j​.livsci​.2009​.01​.016.
cus lactis and Lactobacillus plantarum during pressure treatment. Wang, Y., Z. Ahmed, W. Feng, C. Li, and S. Song. 2008. Physicochemi-
Appl. Environ. Microbiol. 68:4399–4406. https:​/​/​doi​.org/​10​.1128/​ cal properties of exopolysaccharide produced by Lactobacillus kefi-
AEM​.68​.9​.4399​-4406​.2002. ranofaciens ZW3 isolated from Tibet kefir. Int. J. Biol. Macromol.
Nalbantoglu, U., A. Cakar, H. Dogan, N. Abaci, D. Ustek, K. Sayood, 43:283–288. https:​/​/​doi​.org/​10​.1016/​j​.ijbiomac​.2008​.06​.011.
and H. Can. 2014. Metagenomic analysis of the microbial commu-
nity in kefir grains. Food Microbiol. 41:42–51. https:​/​/​doi​.org/​10​
.1016/​j​.fm​.2014​.01​.014. ORCIDS
Ondov, B. D., N. H. Bergman, and A. M. Phillippy. 2011. Interactive
metagenomic visualization in a web browser. BMC Bioinformatics Rafael Jiménez-Flores https:​/​/​orcid​.org/​0000​-0003​-4905​-5021
12:385. https:​/​/​doi​.org/​10​.1186/​1471​-2105​-12​-385.

Journal of Dairy Science Vol. 107 No. 7, 2024