1084 Pereault et al.: Journal of AOAC International Vol. 97, No.
4, 2014
FOOD BIOLOGICAL CONTAMINANTS
Validation of the Soleris® Direct Yeast and Mold Method for
Semiquantitative Determination of Yeast and Mold
in a Variety of Foods
Performance Tested MethodSM 051301
Abstract
A study was carried out to determine the efficacy of the
Soleris® Direct Yeast and Mold (DYM) automated growth-based
method for semiquantitative detection of yeast and mold in a
variety of food products. A probability of detection (POD)
statistical model was used to compare Soleris results at multiple
test thresholds (dilutions) with plate counts determined using
the U.S. Food and Drug Administration Bacteriological
Analytical Manual, Chapter 18, dilution plating procedure.
Fourteen naturally contaminated food products were tested,
with Soleris testing performed at three or more threshold levels
for each food. Using the POD model, the majority of Soleris
test results were in statistical agreement with the reference
plating procedures. The exceptions included a single threshold
level in yogurt, black pepper, dried fruit, and dry pet food, and
two levels in nonfat dry milk and saw palmetto powder. In all
but one of these instances, the exception being pet food, the
statistical disagreement was due to Soleris estimating a higher
level of contamination than the reference method. Results of
ruggedness testing showed that the Soleris method produced
accurate results even when significant variances in a critical
operating parameter, incubation temperature, were introduced.
Results of the internal and independent laboratory validation
studies showed that the Soleris DYM method can be used as an
accurate alternative to conventional dilution plating procedures
for evaluation of yeast and mold counts at threshold levels,
while saving as much as 72 h in analysis time.
Participants
Method Authors
Marcelle Pereault, Susan Alles,
Oscar Caballero, Ron Sarver, Susan McDougal,
Mark Mozola,1 and Jennifer Rice
Neogen Corp., 620 Lesher Pl, Lansing, MI 48912
Reviewers
Yi Chen
U.S. Food and Drug Administration, Center for Food Safety
and Applied Nutrition, 5100 Paint Branch Pkwy, College Park,
MD 20740
Submitted for publication June 10, 2013.
The method was independently tested, evaluated, and certified by
the AOAC Research Institute as a Performance Tested MethodSM. See
http://www.aoac.org/testkits/steps.html for information on certification.
1
Corresponding author’s e-mail: mmozola@neogen.com
DOI: 10.5740/jaoacint.13-181
Joseph Odumeru
Downloaded from https://academic.oup.com/jaoac/article/97/4/1084/5654821 by guest on 13 December 2023
University of Guelph, 95 Stone Rd, Guelph, Ontario H1H 8J7,
Canada
Wayne Ziemer
Consultant, Loganville, GA 30052
Independent Laboratory
Q Laboratories, 1400 Harrison Ave, Cincinnati, OH 45214
Scope of Method
(a) Target organisms.—Yeasts and molds.
(b) Matrixes.—The following foods were tested: nonfat dry
milk, ice cream mix, salad dressing, yogurt, dried fruit, orange
juice concentrate, tomato juice, saw palmetto powder, corn
flour, cocoa powder, cocoa butter, cocoa liquor, dry pet food,
and black pepper.
(c) Summary of validated performance claims.—Statistically
equivalent to the U.S. Food and Drug Administration
Bacteriological Analytical Manual (FDA/BAM) Chapter 18,
dilution plating method (1) for enumeration of yeasts and molds
in foods at threshold levels using a probability of detection (POD)
model.
Definitions
(a) POD.—The proportion of positive analytical outcomes
for a qualitative method for a given matrix at a given analyte
level or concentration. POD is concentration dependent.
(b) POD (obs.).—Observed POD for the test method;
number of test method positive results divided by the number of
portions tested, with 95% confidence interval.
(c) POD (pred.).—POD for the test method predicted from
the reference method result. POD (pred.) = 1 – e–c, where e is
the base of the natural logarithm (value = 2.7183) and c is the
number of input CFU per test based on the reference method
result.
(d) Pass/Fail result.—If the POD (pred.) falls within the
lower and upper 95% confidence limits of the POD (obs.), the
result is Pass, otherwise Fail.
(e) Soleris test threshold.—The detection limit of the Soleris
method determined by the volume and dilution of test sample
homogenate added to the Soleris vial. For example, a threshold
of >100 CFU/g represents 1 mL of a 1:100 dilution added to the
vial, equivalent to 0.01 g of original test sample.
 Pereault et al.: Journal of AOAC International Vol. 97, No. 4, 2014 1085
Principle
Soleris® Direct Yeast Mold (DYM) is a growth-based,
automated method with an optical endpoint (2). A dilution of
the test sample homogenate is inoculated directly into the test
vial. The prepared vial is then placed in the Soleris instrument
set at the appropriate incubation temperature. As yeasts and
molds grow and metabolize, carbon dioxide is generated, which
diffuses from the growth medium through a gas-permeable
layer and into the indicator portion of the Soleris vial. Dissolved
carbon dioxide leads to the formation of carbonic acid, reducing
the pH and resulting in a change in color of the chemical
indicator over time. The color is monitored by the instrument,
and a change in color of a certain magnitude, determined by
the Soleris software, is indicative of a positive test result. The
Soleris instrument contains temperature-controlled incubation
chambers and photodiode-based optical detection devices for
measurement of color changes in the bottom indicator portion
of the vial.
The Soleris DYM test is used in a “dilute-to-specification”
or threshold manner, in which the result is positive or negative
around a desired cutoff (in CFU/g) determined by the dilution
and volume of sample homogenate added to the vial. It is
assumed that 1 CFU introduced into the Soleris vial will lead
to a positive result.
General Information
Standard reference methods for detection and enumeration of
yeasts and molds in foods include the FDA/BAM dilution plating
technique and similar methods. These methods typically require
5 days to obtain results. The Soleris DYM method, through
sensitive determination of metabolic activity during microbial
growth, produces results within 48 h for most foods. The Soleris
DYM vial represents an improvement over the original Soleris
yeast and mold vial. While the original vial detected via carbon
dioxide as well, this was done in a “vial-in-vial” format, in
which an inner vial containing selective growth medium and
a growth substrate was inoculated and placed (without a cap)
into an outer vial containing a liquid pH indicator in the optical
reading window at the bottom. This method featured 72 h test
duration and was validated for a variety of foods (apple pie, dry
pet foods, nonfat dry milk, ranch salad dressing, saw palmetto,
tomato juice, hard salami, soy flour, and strawberry puree).
SM
The test was granted Performance Tested Method status
(No. 040901) in 2009 (3). In the new DYM format, detection
of carbon dioxide occurs directly in a single vial, making it a
more efficient detection process. This increased efficiency, in
turn, leads to significantly reduced detection times.
Materials and Methods
Test Kit Information
®
(a) Test name.—Soleris DYM, 9 mL.
(b) Cat. No.—DYM-109.
(c) Ordering information.—Inside the United States.—
Neogen Corp., 620 Lesher Pl, Lansing, MI 48912, Tel: 517-372-
9200, Fax: 517-372-0108, Web site: www.neogen.com. Outside
the United States.—Contact above office for local distributor
information.
(d) Soleris DYM vial.—Sterile growth medium (9 mL) in
a plastic vial, in boxes of 100 vials. One test sample per vial.
Store at 2–8°C. Requires use of the Soleris instrument.
Supplies and Reagents
Additional supplies and reagents available from Neogen
Corp.
(a) Yeast and Mold Supplement (Product no. YI-110).—
Four vials/package, 10 mL each. Store at 2–8°C.
(b) Soleris 32 (product No. BSX32) or Soleris 128
(product No. BSX128) instrument.—Containing one or four
temperature-controlled (15–60 ± 0.5°C) incubator drawers,
Downloaded from https://academic.oup.com/jaoac/article/97/4/1084/5654821 by guest on 13 December 2023
respectively, with 32 locations in each drawer. Each test location
contains an LED-based optical sensor for measurement of
changes in absorbance over time. The system includes a dedicated
computer, monitor, and software.
Other required supplies
®
(c) Stomacher -type bags.—With filter (product No. 6827).
(d) Balance.—For weighing samples, minimum 100 g
capacity ±0.1 g.
(e) Micropipettor and tips.—100 to 1000 µL.
(f) Ethanol.—200 proof.
Safety Precautions
Use of this test should be restricted to individuals with
appropriate laboratory training in microbiology. Reagents are
for laboratory use only. All pipetting transfers should be made
using either a disposable pipet and pipetting aid or micropipet
with disposable tips. Refer to the Material Safety Data Sheet
available from Neogen Corp. for more information. Used Soleris
vials, sample homogenates and dilutions, and pipets or pipet
tips should be handled and disposed of as potentially infectious
material. The preferred method for disposal of contaminated
materials is autoclaving. Items that cannot be autoclaved may
be decontaminated by treatment with a disinfectant solution, for
example 10% bleach, followed by rinsing with water.
Preparation of Yeast & Mold Supplement
(a) To one vial of Yeast and Mold Supplement, add 1.0 mL of
200 proof ethanol. Wet powder well.
(b) Add 9 mL sterile deionized water. Mix well. Store in
refrigerator up to 7 days after rehydration.
Sample Preparation
(a) Prepare a 1:10 sample homogenate by adding 225 mL
0.1% peptone water to 25 g food sample. Homogenize the
sample in a Stomacher for 2 min.
(b) Prepare dilutions of the sample homogenate to achieve
the desired test threshold, e.g., for >10 CFU/g, use sample
homogenate as is; for >100 CFU/g, prepare further 1:10 dilution
in 0.1% peptone water; for >1000 CFU/g, prepare further
1:100 dilution, etc.
(c) Add rehydrated Yeast and Mold Supplement to each vial,
0.60 mL for products containing a bacterial starter culture, or
0.15 mL for all other foods.
(d) Add 1.0 mL of sample homogenate or dilution to the
1086 Pereault et al.: Journal of AOAC International Vol. 97, No. 4, 2014
Soleris vial. Cap the vial tightly and invert three times to mix.
Back off the vial cap to allow air exchange.
(e) Proceed to Soleris analysis.
General Preparation
The test should be performed under normal laboratory
conditions with respect to humidity, temperature, lighting, etc.
Soleris vials should not be used beyond their expiration date.
Soleris Analysis
Note: The Soleris system requires installation and user
training. Both are provided by Neogen Corp.
(a) In the Soleris software menu, select an incubator drawer.
The incubation temperature should be set to 28 ± 0.5°C for all
foods.
(b) On the sample position grid, select the test (DYM) and
drag/drop to the selected sample position.
(c) On the sample queue screen, enter the following: Sample
identification number, and, if desired, production lot number,
plant information, and user name.
(d) Place the inoculated Soleris vials into the selected drawer
locations.
(e) Once all the samples are entered, select and click Start.
(f) Incubate test samples for 48 h. A detection curve will
be generated in real time. The Soleris software will indicate a
positive test result. Positive results will generally be indicated in
less than 48 h. Note: If testing cocoa butter, set the test duration
to 60 h.
Interpretation of Results
(a) Negative criterion.—Tests producing no detection within
48 h are considered negative at the test threshold selected.
(b) Positive criterion.—Detection times within 48 h verified
by visual confirmation of indicator color change (blue-green to
yellow-green or yellow) indicates a positive result at the test
threshold selected.
Internal Validation Studies
Comparative Testing of Naturally Contaminated or
Inoculated Foods
The Soleris DYM method was compared to the FDA/BAM
directing plating technique for all foods tested. Naturally
contaminated foods were used where possible. However,
inoculation was required in most cases to achieve sufficiently
high levels for testing at multiple threshold levels. Inoculation
of test products was performed using pure cultures from the
American Type Culture Collection (ATCC; Manassas, VA) or
other documented sources (Table 1), or in some cases, using
natural yeast or mold flora isolated from similar product types.
Preparation of inocula.―Molds were grown on potato
dextrose agar (PDA) plates for 5 days at 25°C or until plates
were completely dry. Yeast strains were grown in tryptic soy
broth with yeast extract for 24–48 h at 30°C. Procedures for
preparation of mold inocula were adapted from those described
by Entis and Lerner (4).
Low-moisture products (nonfat dry milk, dried fruit, saw
palmetto powder, corn flour, cocoa powder, cocoa liquor, cocoa
butter, dry pet food, and black pepper).—All low-moisture
products, except nonfat dry milk and cocoa butter, were
inoculated with mold. Nonfat dry milk and cocoa butter were
inoculated with yeast. For dry products, the mold culture
plate was placed in a plastic bag with approximately 1 kg of
test product and shaken. After equilibration for a minimum of
14 days at room temperature, the level of contamination was
determined by plate count and the seeded product diluted with
additional test product, if necessary, to achieve desired levels.
Further details of sample preparation for individual foods
follow.
Nonfat dry milk was inoculated with Saccharomyces
Downloaded from https://academic.oup.com/jaoac/article/97/4/1084/5654821 by guest on 13 December 2023
cerevisiae using the following procedure. A liquid suspension
of the yeast culture was prepared and dried in a Petri dish.
Dried cells were then scraped up and used to inoculate the test
product. Cocoa butter was melted, inoculated with Rhodotorula
glutinis, and allowed to re-solidify. Cocoa liquor was melted,
inoculated with spores of Aspergillus ochraceus, and allowed
to re-solidify. Corn flour was inoculated with a sporulated mold
plate obtained from a naturally contaminated corn cob. Cocoa
powder was inoculated with a natural mold isolate from the
same product. This isolate was cultured to sporulation and then
used to inoculate the product at a higher level.
All other products were inoculated with mold pure cultures as
described above (see Table 1 for inoculum strains).
High-moisture products (ice cream mix, ranch salad dressing,
yogurt, orange juice concentrate, and tomato juice).—Products
were inoculated with an appropriate dilution of an overnight
culture of the yeast test strain. For products inoculated with
molds, growth was harvested from the culture plate and a
spore suspension was prepared in buffer (Maximum Recovery
Diluent or equivalent). The spore titer was determined using
a hemocytometer. An aliquot of an appropriate dilution of
the spore preparation was added to the product and mixed
thoroughly. Details of sample preparation for individual foods
follow.
Ice cream mix was inoculated with a mixed yeast culture
representing natural contaminants from the same product in
order to achieve higher levels. A low-level natural contaminant
was found in orange juice concentrate. Product was abused for
2 days at 25°C and then re-frozen. Product was subsequently
thawed and tested. For tomato juice, a mold was isolated from
a fresh tomato and cultured as described above. This mold
preparation was then used to inoculate the product. Ranch salad
dressing and yogurt were inoculated with pure cultures using
procedures described above (see Table 1 for inoculum strains).
All inoculated products except orange juice concentrate were
held for 48–72 h at 2–8°C for equilibration of the inoculum
prior to testing. Orange juice concentrate was held at –20°C for
14 days.
Sample preparation.—For each food type, a 1:10 sample
homogenate (25 g plus 225 mL buffer) was prepared in
accordance with the reference method. Further dilutions were
prepared for Soleris testing at three or more test thresholds and
for the reference method plate count. Soleris test thresholds
were chosen based on an estimate of the microbial load of the
test sample and with the goal of achieving fractional positive
results for at least one level, i.e., a target of 5–15 positive results
out of 20 replicate assays at a particular dilution. This ensures
that the Soleris method has been challenged at an inoculation
Table 1. Comparative testing results and probability of detection calculations for the Soleris DYM method
Soleris result Observed PODf

Soleris test
Reference plate threshold Predicted No. vials No. vials
a b c d e g
Food type Inoculum/naturally contaminated count (CFU/g) (CFU/g) CFU/vial POD positive tested LCI POD UCI Interpretation

Ranch dressing Zygosaccharomyces bailii ATCC 42476 1.26 × 105 >10000 12.6 1.00 20 20 0.83 1.00 1.00 Pass

>100000 1.26 0.72 13 20 0.43 0.65 0.82 Pass

>1000000 0.126 0.12 0 20 0 0 0.17 Pass

<10 >10 0 0 0 20 0 0 0.17 Pass

4
Yogurt Candida famata ATCC 60229 4.12 × 10 >1000 41.2 1.00 20 20 0.83 1.00 1.00 Pass

>10000 4.12 0.98 20 20 0.83 1.00 1.00 Pass

>100000 0.412 0.34 12 20 0.39 0.60 0.78 Fail

<10 >10 0 0 0 20 0 0 0.17 Pass

Yogurth Candida famata ATCC 201067 1.20 × 102 >10 12.0 1.00 20 20 0.83 1.00 1.00 Pass

>100 1.20 0.70 14 20 0.48 0.70 0.85 Pass

>1000 0.120 0.11 7 20 0.18 0.35 0.57 Fail

<10 >10 0 0 0 20 0 0 0.17 Pass

Ice cream mix Nat. contam. yeast 1.25 × 105 >10000 12.5 1.00 20 20 0.83 1.00 1.00 Pass

>100000 1.25 0.71 15 20 0.53 0.75 0.89 Pass

>1000000 0.125 0.12 4 20 0.08 0.20 0.42 Pass

1.93 × 103 >10 193 1.00 20 20 0.83 1.00 1.00 Pass

4
Nonfat dry milk Saccharomyces cerevisiae 4.42 × 10 >10000 4.42 1.00 20 20 0.83 1.00 1.00 Pass

>100000 0.442 0.36 12 20 0.39 0.60 0.78 Fail

>1000000 0.0442 0.04 4 20 0.08 0.20 0.42 Fail

<10 >10 0 0 0 20 0 0 0.17 Pass

Orange juice concentrate Nat. contam. yeast 1.05 × 106 >100000 10.5 1.00 20 20 0.83 1.00 1.00 Pass

>1000000 1.05 0.65 12 20 0.39 0.60 0.78 Pass

>10000000 0.105 0.10 4 20 0.08 0.20 0.42 Pass

>100000000 0.0105 0.01 0 20 0 0 0.17 Pass

<10 (5)i >10 0.5 0.39 7 20 0.18 0.35 0.57 Pass

Tomato juice Nat. contam. mold 9.35 × 104 >10000 9.35 1.00 20 20 0.83 1.00 1.00 Pass

>100000 0.935 0.61 14 20 0.48 0.70 0.85 Pass

>1000000 0.0935 0.09 3 20 0.05 0.15 0.36 Pass

<10 >10 0 0 0 20 0 0 0.17 Pass

Pereault et al.: Journal of AOAC International Vol. 97, No. 4, 2014 1087

Downloaded from https://academic.oup.com/jaoac/article/97/4/1084/5654821 by guest on 13 December 2023

Table 1. (continued)
1088

Soleris result Observed PODf

Soleris test
Reference plate threshold Predicted No. vials No. vials
a b d e
Food type Inoculum/naturally contaminated count (CFU/g) (CFU/g) CFU/vialc POD positive tested LCI POD UCI Interpretationg
h 1
Tomato juice Penicillium chrysogenum ATCC 10106 1.70 × 10 >10 1.70 0.82 14 20 0.48 0.70 0.85 Pass

>100 0.170 0.16 1 20 0.01 0.05 0.24 Pass

>1000 0.017 0.02 0 20 0 0 0.17 Pass

<10 >10 0 0 0 20 0 0 0.17 Pass

Dried fruit Mucor racemosus ATCC 42647 3.78 × 104 >1000 37.8 1.00 20 20 0.83 1.00 1.00 Pass

>10000 3.78 0.98 19 20 0.76 0.95 0.99 Pass

>100000 0.378 0.31 6 20 0.14 0.30 0.52 Pass

>1000000 0.0378 0.04 1 20 0.01 0.05 0.24 Pass

<10 >10 0 0 1 20 0.01 0.05 0.24 Fail

4
Corn flour Nat. contam. mold 8.21 × 10 >10000 8.21 1.00 20 20 0.83 1.00 1.00 Pass

>100000 0.821 0.56 14 20 0.48 0.70 0.85 Pass

>1000000 0.0821 0.08 2 20 0.03 0.10 0.30 Pass

<10 (4)i >10 0.4 0.33 9 20 0.26 0.45 0.66 Pass

5
Cocoa powder Nat. contam. mold 5.30 × 10 >100000 5.30 1.00 20 20 0.83 1.00 1.00 Pass

>1000000 0.530 0.41 11 20 0.34 0.55 0.74 Pass

>10000000 0.0530 0.05 3 20 0.05 0.15 0.36 Pass

33 >10 3.30 0.96 20 20 0.83 1.00 1.00 Pass

h 5
Pereault et al.: Journal of AOAC International Vol. 97, No. 4, 2014

Cocoa powder Aspergillus fumigatus ATCC 204305 5.00 × 10 >10000 50.0 1.00 20 20 0.83 1.00 1.00 Pass

>100000 5.00 0.99 19 20 0.76 0.95 0.99 Pass

>1000000 0.500 0.39 7 20 0.39 0.60 0.78 Pass

<10 >10 0 0 0 20 0 0 0.17 Pass

5
Cocoa liquor Aspergillus ochraceus ATCC 1008 5.30 × 10 >100000 5.30 0.99 20 20 0.83 1.00 1.00 Pass

>1000000 0.530 0.41 5 20 0.11 0.25 0.47 Pass

>10000000 0.0530 0.05 1 20 0.01 0.05 0.24 Pass

<10 >10 0 0 0 20 0 0 0.17 Pass

3 j
Cocoa butter Rhodotorula glutinis ATCC 15125 8.94 × 10 >1000 8.94 1.00 20 20 0.83 1.00 1.00 Pass

>10000 0.894 0.59 9j 20 0.26 0.45 0.66 Pass

j
>100000 0.0894 0.09 3 20 0.052 0.15 0.360 Pass

<10 >10 0 0 0 20 0 0 0.17 Pass

Downloaded from https://academic.oup.com/jaoac/article/97/4/1084/5654821 by guest on 13 December 2023

Table 1. (continued)
Soleris result Observed PODf

Soleris test
Reference plate threshold Predicted No. vials No. vials
a b d e
Food type Inoculum/naturally contaminated count (CFU/g) (CFU/g) CFU/vialc POD positive tested LCI POD UCI Interpretationg
5
Black pepper Aspergillus tamari ATCC 26950 4.00 × 10 >100000 4.00 1.00 20 20 0.83 1.00 1.00 Pass

>1000000 0.400 0.33 10 20 0.30 0.50 0.70 Pass

>10000000 0.0400 0.04 3 20 0.05 0.15 0.36 Fail

i
<10 (1) >10 0.1 0.09 1 20 0.01 0.05 0.24 Pass
Saw palmetto powder Aspergillus niger ATCC 16404 1.31 × 104 >1000 13.1 1.00 20 20 0.83 1.00 1.00 Pass

>10000 1.31 0.73 20 20 0.83 1.00 1.00 Fail

>100000 0.131 0.12 9 20 0.26 0.45 0.66 Fail

>1000000 0.0131 0.01 0 20 0 0 0.17 Pass

<10 >10 0 0 0 20 0 0 0.17 Pass

5
Dry pet food Fusarium graminearum ATCC MYA-932 1.03 × 10 >10000 10.3 1.00 20 20 0.83 1.00 1.00 Pass

>100000 1.03 0.64 7 20 0.18 0.35 0.57 Fail

>1000000 0.103 0.10 1 20 0.01 0.05 0.24 Pass

<10 >10 0 0 0 20 0 0 0.17 Pass

a
 Direct plating method: Bacteriological Analytical Manual, Chapter 18, dilution plating method.
b
 Dilute-to-specification approach. Positive result expected if organism concentration in test dilution is higher than Soleris test threshold.
c
 CFU per Soleris vial based on reference plate count.
d
 POD (predicted): Based on reference plate count (POD = 1 – e–c).
e
 Detection times within 48 h indicate a positive result at the test threshold selected.
f
 POD (observed): Fraction of Soleris results positive. LCI and UCI are 95% lower and upper confidence intervals, respectively.
g
 Test for equivalence of reference plate count and Soleris results. For “Pass”, POD (pred.) must lie with POD (obs.) LCI-UCI range.
h
 Trial performed by independent laboratory.
i
 Reference method count below 10 CFU/g possibly due to use of 10 plates.
j
 Test duration of 60 h.
Pereault et al.: Journal of AOAC International Vol. 97, No. 4, 2014 1089

Downloaded from https://academic.oup.com/jaoac/article/97/4/1084/5654821 by guest on 13 December 2023

1090 Pereault et al.: Journal of AOAC International Vol. 97, No. 4, 2014
Table 2. Ruggedness testing results for the Soleris DYM method
5.3
Incubation temp., °C No. positivesa Mean detection time, h
26 9/10 45.2
28 9/10 42.6
30 10/10 41.5
a
 Number of vials positive within 48 h.
b
 No detection within 48 h.
level of approximately 1 CFU per vial or less. A homogenate
was also prepared from uninoculated test material, and this
homogenate was tested at the >10 CFU/g threshold level.
Soleris testing.—For the Soleris method, 20 vials were
inoculated with 1 mL of dilution for each test threshold.
Duration of the Soleris test was 48 h for all foods, except 60 h
for cocoa butter.
Reference method plate counts.—The FDA/BAM dilution
plating method was performed using dichloran rose bengal
chloramphenicol agar (or dichloran 18% glycerol agar for
foods with aw < 0.95). The method was modified by plating
10 replicates of each dilution, rather than the specified triplicate
platings, in order to obtain a more accurate count. Standard
counting rules were followed, i.e., counts were based on plates
containing 30–300 colonies whenever possible.
Data analysis.―Results for each food type were analyzed
using a POD model based on the Poisson distribution. The
predicted POD for the Soleris method was calculated using
the formula POD (pred.) = 1 – e–c, where c = CFU/vial at a given
test threshold based on the reference method plate count. The
observed POD for the Soleris method was calculated at each
test threshold by dividing the number of positive Soleris results
by the number of portions tested (20). Ninety-five percent
confidence intervals around POD (obs.) were calculated using a
POD interval calculator (P. Wehling, personal communication).
If POD (pred.) was within the confidence interval of POD
(obs.), the Soleris and reference method results were considered
to be statistically not different.
Example
Reference method result.—63 CFU/g.
Soleris test threshold.—>100 CFU/g.
Soleris POD (pred.).—1 – e–0.63 = 0.467
Soleris POD (obs.).—9/20 = 0.450
Soleris POD (obs.) with confidence intervals.—0.258–0.658
Result.—Pass
Results.—Results are reported in Table 1. FDA/BAM direct
method plate counts for abused or inoculated test products
ranged from a low of 8.9 × 103 CFU/g for dried fruit, to a high
of 1.0 × 106 CFU/g for orange juice concentrate. Most test
products, without abuse or inoculation, produced plate counts
of <10 CFU/g. Ice cream mix was naturally contaminated at a
level of 1.9 × 103 CFU/g. Orange juice concentrate, corn flour,
cocoa powder, and black pepper all showed very low levels of
natural contamination, ranging from 1 to 33 CFU/g. Detection
of counts below 10 CFU/g was facilitated by the use of 10 plates
in the reference method analyses.
Input level (CFU/vial)
0.53 0
No. positives Mean detection time, h No. positives Mean detection time, h
b
0/10 ND 0/5 ND
2/10 45.2 0/5 ND
1/10 47.0 0/5 ND
Downloaded from https://academic.oup.com/jaoac/article/97/4/1084/5654821 by guest on 13 December 2023
All trials produced fractional positive results at one or more
test threshold levels with the Soleris method. Based on POD
analysis, there were no statistically significant differences in
results between the Soleris and reference methods at any test
threshold for eight of 14 commodities. For three of the remaining
foods, yogurt, dried fruit, and black pepper, there was one test
threshold at which there was a statistically significant difference.
This was due to Soleris estimating a slightly higher level of
contamination than the reference method. For the two remaining
foods, nonfat dry milk and saw palmetto powder, this occurred
at two threshold levels. There was only one commodity, dry pet
food, at the >100 000 CFU/g test threshold, where the reference
method indicated a higher level of contamination, with POD
(pred.) being slightly outside the POD (obs.) upper confidence
limit (0.64 versus 0.57). The dry pet food data, however, passed
the statistical test at all other threshold levels tested.
For uninoculated control samples, Soleris and reference
method results were in statistical agreement for all foods except
dried fruit. For this matrix, Soleris produced one positive vial
while the reference method plate count was <10 CFU/g.
For cocoa butter only, it was determined that a test duration
of 48 h was insufficient, and that a test duration of 60 h is
required for maximum test sensitivity with this matrix. At the
intermediate dilution, there were only five positive vials at 48 h
(failing the POD test) versus nine positive vials at 60 h. At the
lowest dilution, there were zero and three positive vials at 48
and 60 h, respectively.
Ruggedness Testing
Methodology.—Soleris is an automated microbiology
platform with instrument settings determined by the software
specific to applications such as the DYM test. Incubation
temperature is controlled to a tolerance of ± 0.5°C. To test
method ruggedness beyond these limits, tests were performed
in which instruments were set to incorrect incubation
temperatures, specifically 26 and 30°C versus the specified
28°C. Test samples included dilutions of inoculated cocoa
liquor prepared to contain approximately 5 and 0.5 CFU/mL
of Aspergillus ochraceus. Uninoculated cocoa liquor samples
were used as negative controls.
Results.—Results are shown in Table 2. For both dilutions
tested, there were similar numbers of positive vials at all three
temperatures tested. At the 5.3 CFU/vial input level, mean
detection times for vials incubated at 26, 28, and 30°C were
45.2, 42.6, and 41.5 h, respectively. The number of vials positive
at the three temperatures was also similar for the 0.53 CFU/vial
input level. Even this relatively large variation in incubation
 Pereault et al.: Journal of AOAC International Vol. 97, No. 4, 2014 1091
temperature failed to produce a demonstrable negative effect on
the Soleris results. In practice, a temperature variance of this
magnitude would not be encountered in normal operation of the
Soleris instrument.
Independent Laboratory Study
Methodology.—The independent laboratory tested three
foods: yogurt, tomato juice, and cocoa powder. Each matrix was
artificially contaminated with a single yeast or mold strain at a
target level of 100–1 000 000 CFU/g. For yogurt, the Candida
famata culture was prepared by transferring a single yeast colony
from trypticase soy agar with 5% sheep’s blood to trypticase
soy broth and incubating the culture for 48 ± 2 h at 30 ± 1°C.
The inoculum was prepared by serially diluting the culture in
0.1% peptone water. An aliquot of diluted culture was used
to artificially contaminate a bulk sample of yogurt. Following
inoculation, the bulk sample of yogurt was held at 2–5°C for
48–72 h to allow the organism to equilibrate within the matrix.
The Penicillium chrysogenum (ATCC 10106) and Aspergillus
fumigatus (ATCC 204305) cultures were prepared on PDA and
incubated for 7 days at 25 ± 1°C. For the tomato juice inoculum,
a spore suspension of P. chrysogenum was created by washing
the culture plates with 20 mL of wash buffer (0.9% saline and
0.05% Tween 80). The spore suspension was serially diluted in
0.1% peptone water. An aliquot of diluted spore suspension was
used to artificially contaminate a bulk sample of tomato juice.
Following inoculation, the bulk sample of tomato juice was
held at 2–5°C for 48–72 h to allow the organism to equilibrate
within the matrix. For cocoa powder, the mold culture plate of
A. fumigatus was placed inside a Stomacher bag containing 1 kg
of test product. The Stomacher bag was shaken vigorously and
stored at room temperature (24 ± 2°C) for 14 days to allow the
organism to equilibrate within the matrix. Background screens
of the inoculated products were performed to estimate the
microbial load of the test products and determine the dilution
expected to yield fractional positive results with the Soleris
method. All other details of the analysis were in accordance
with the procedures described above for the internal trials.
Results.—Results are shown in Table 1. Results of the Soleris
and reference methods were in statistical agreement for all three
foods at all test threshold levels, with the single exception of
yogurt at the >1000 threshold level, where the Soleris method
determined a slightly higher level of contamination [POD
(pred.) of 0.11 versus POD (obs.) lower confidence interval of
0.18]. Overall, results of the independent laboratory trials were
consistent with those of the internal studies.
Discussion
The results presented in this study demonstrate that the Soleris
DYM method can be used as an effective alternative to the
FDA/BAM reference plating procedure. With few exceptions,
the Soleris method results were statistically equivalent to
those of the reference plating procedure as determined using
a POD model. Where there was lack of statistical equivalence,
in most cases the Soleris method estimated a higher level of
Downloaded from https://academic.oup.com/jaoac/article/97/4/1084/5654821 by guest on 13 December 2023
contamination (eight of nine instances; see Table 1). Use of
the Soleris DYM method in a “dilute to specification” mode
allows users to match test thresholds with product release
specifications. Compared to direct plating procedures, the
Soleris method offers the advantages of minimal labor and
reduced analysis time; results are obtained in 48 h or less for
most foods, compared to the 5 days required by conventional
plating methods.
Conclusions
It is recommended that the Soleris DYM method be granted
Performance Tested Method status for estimation of yeasts
and molds at designated threshold levels in nonfat dry milk,
ice cream mix, salad dressing, yogurt, dried fruit, orange juice
concentrate, tomato juice, saw palmetto powder, corn flour,
cocoa powder, cocoa liquor, cocoa butter, dry pet food, and
black pepper.
References
(1) U.S. Food and Drug Administration (2001) Bacteriological
Analytical Manual Online, Chapter 18, Yeasts, Molds, and
Mycotoxins http://www.fda.gov/Food/FoodScienceResearch/
LaboratoryMethods/ucm071435.htm
(2) Neogen Corp. (2011) Soleris Operator’s Manual, Version 7,
Lansing, MI
(3) Alles, S., Shrestha, N., Ellsworth, A., Rider, A., Foti, D.,
Knickerbocker, J., & Mozola, M. (2009) J. AOAC Int. 92,
1396–1415
(4) Entis, P., & Lerner, I. (1996) J. Food Prot. 59, 416–419