About the Program
The Office of the Texas State Chemist (OTSC) – Texas A&M AgriLife Research Aflatoxin Proficiency Testing and Control program includes sending quarterly
proficiency samples of aflatoxin (5 – 300 ppb).
In addition, you can also register for bi-annual samples of fumonisin (0.5 – 30 ppm) and/or deoxynivalenol (0.5 – 30 ppm).
Please note that the aflatoxin sample should be used only for aflatoxin analysis, and the fumonisin and DON sample
should be used only for fumonisin and DON analysis respectively. In particular, 100+ g sample(s) will be mailed by OTSC twice a year.
For each round, participants are expected to analyze duplicate 50 g portions and report both results to OTSC
via the web portal, at their own expense for each mycotoxin. You can register for your choice of analytes by
updating your information on the Update Contact Information page.
Eligible participants include grain milling and grain handling firms, feed manufacturing industry, government laboratories
and educational institutions. The total number of participants is limited to 300 participants. Results will be reviewed for
outliers before the participant mean, range, and standard deviation of the results are calculated. The z-score for each laboratory
will be calculated and reported. OTSC will not divulge the identity of program participants.
Maize Samples
The maize samples used for the program contain naturally occurring aflatoxin, fumonisin or deoxynivalenol. The samples containing aflatoxin
are analyzed for aflatoxin B1, B2, G1, and G2 by OTSC. Similarly, the samples containing fumonisin are analyzed for total
fumonisin (FB1 + FB2 + FB3) by OTSC. Samples are analyzed using high performance liquid chromatography (aflatoxin) or LC-MS/MS (fumonisin and deoxynivalenol),
and the relative standard deviations for the proficiency samples are calculated prior to their incorporation into the proficiency program.
The Office of the Texas State Chemist is ISO/IEC 17025:2017 accredited for aflatoxin and fumonisin analysis using these procedures and
ISO 17034:2016 and ISO/IEC 17043:2010 accredited for conducting proficiency testing and producing reference material, respectively.
Results Analysis
The PT sample production follows the ISO/IEC 17043:2010 standard, and the final products
have passed a homogeneity test.
Each PT item has an assigned value (µ). This assigned value is based on the result
from in-house homogeneity testing. The z-score calculation is based on the
following equations:
Z = (x - µ) / σ
where Z is the z-score to evaluate the individual laboratory performance, X is the mean value of the
two individual laboratory reported numbers, and µ is the assigned value for the corresponding PT
item. σ is calculated using the Horwitz equation to initially determine the predicted relative standard deviation (RSD):
RSD(%) = 2(1 -
log(C)
2
)
where C is the concentration of aflatoxin expressed as a dimensionless mass fraction. (For example, if the
concentration is in ppb, then the number that goes into the equation would be C×10-9.)
σ =
RSD(%)
100
* µ
The assigned standard deviation is σ.
Z-score evaluation:
Correction: A correction is taken to rectify a known non-conformance. In the case of a questionable z-score,
the lab should take a look at the result and determine if there are any changes that need to take place to improve quality.
For example, instrument maintenance, expiry of reagents, pipet technique, analyst training, etc.
Corrective action: A corrective action is taken to prevent recurrence of said nonconformance. With an unsatisfactory z-score,
a correction may only be a temporary solution. A corrective action delves deeper and seeks to find a root cause for the nonconformance.
A correction can get a lab back up and running, but a corrective action may save a lab lots of work in the future.
An additional event that may require a correction and/or corrective action is when lab results are identified as a Cochran and/or
Grubbs Outlier (see section on ‘Outliers’ for more information).
- It is possible to have a “satisfactory z-score”, but be identified as a Cochran outlier. Cochran identifies results that
vary greatly between duplicates (as compared to the inherent variability between duplicates within the entire data set).
- In this PT scheme, because an assigned mean and assigned standard deviation are used, Grubbs outliers will always be
associated with an unsatisfactory z-score.
Outlier Test
The outlier test is based on Cochran C test and Grubbs test. The calculation will
be based on the following equation:
Where Ȼj is Cochran’s C statistic for data series j, sj is standard deviation of
an individual laboratory’s results from the assigned mean, and si is the standard deviation for each laboratory that
submitted a result for this round. The highest Ȼj is then compared to a standard Cochran number based on the number
of labs participating. If the highest Ȼj is bigger than the standard Cochran number, that laboratory is an outlier
and removed from the set. The Cochran statistic is now calculated for the labs remaining in the set. The process
continues as stated above until no other outliers are found.
Next, the Grubbs Outlier test is run on the dataset with the Cochran outliers taken out. The Grubbs’ test can be used to
determine whether the maximum value (or minimum value) is an outlier. The Grubbs test statistic is calculated
based on the following equation:
Where Y is the sample mean, and s is the standard deviation. The Grubbs
statistic is compared to a standard Grubbs Number (based on number of results submitted), if the G statistic is higher than the Grubbs
number, the lab with the highest difference is removed from the set. The Grubbs statistic is now calculated for the labs remaining in
the set. The process continues as stated above until no other outliers are found. (Additional information on the Cochran C Test and
Grubbs Test can be found here: www.rsc.org/images/TB%2069_tcm18-247571.pdf)
Sample Analysis
For each round, participants are expected to analyze duplicate 50 g portions and report both results to OTSC via the web portal,
at their own expense for aflatoxin, fumonisin and deoxynivalenol respectively respectively. The testing of the received PT item should follow the same format
of routine analysis of laboratory samples. The PT samples are stable in the course of production, transportation, PT testing and
reporting period. The PT item is ready to use in its delivered form and does not require any extra laboratory preparation. Typical
laboratory environmental conditions are sufficient to perform the expected measurements. The PT item is the same as normal maize
samples and does not require special safety handling procedures.
Return of the PT sample is not required for this program.
Test Methods
The participating laboratories can use any method of choice. The chosen method must be provided with the results.
Eligible methods include, but are not limited to, traditional instrument analysis (HPLC) and testing kit methods.
The data analysis for results obtained by different methods will be reported in three categories: 1) testing kits,
2) other methods, and 3) all methods combined.
Reporting Results
The testing results are reported through the Proficiency Testing website: pt.tamu.edu/LoginRequired/DataEntry.aspx.
Each laboratory participant representative shall use the lab number issued by the PT provider (OTSC) and log into the
“Laboratory Data Reporting” section of the APTECA website. Two independent analyses (weighing, extraction and analysis)
shall be performed for the same PT item and two results are reported. For aflatoxin, results can be reported for total
aflatoxin or aflatoxin B1 or both. For fumonisin, results are reported as total fumonisin. In addition, the laboratory
method of choice will need to be reported.
The final reporting shall not be later than the date specified in the PT scheme,
which is announced on the program website (pt.tamu.edu).
PT Summary Report Usage Policy
The laboratory participant can use the PT summary report for their legitimate purpose when
crediting this PT program. OTSC-TAMU Agrilife reserves the rights to use the data for analysis
and publication purposes. OTSC will not divulge the identity of program participants.
PT Results Appeal Policy and Procedures
The participant laboratory shall appeal the PT results by contacting the PT program manager via written request (via e-mail at apteca@otsc.tamu.edu).
The PT provider will investigate the appeal by arranging a three-member committee, which shall be composed of at least
two members who are not affiliated with the PT provider’s organization. The committee shall determine the validity of the PT results or a new PT item shall be issued for testing.
Contact Information
Tim Herrman, PhD
Professor, State Chemist and Director
Office of the Texas State Chemist
Texas A&M AgriLife Research
P.O. Box 3160
College Station, TX 77841-3160
apteca@otsc.tamu.edu