*4.3. Statistical Analysis*

Descriptive and inferential data analyses were done on statistical computer program (IBM SPSS Statistics 20). For spiking and recovery data, bias was calculated accordingly: %bias = mean% recovery-100. AFB1 recovery of the conventional dry milling procedure and the novel slurry method were analyzed by comparing means of Groups I and II employing independent *t*-test. Variability expressed as variance, standard deviation and coefficient of variation (relative standard deviation) for Groups I and II were computed and compared. Means of all wet milling procedures recovery data (Groups III–V) were compared employing Welch ANOVA to determine between and within the group variation. Games–Howell test (a post-hoc test) incorporated for multiple comparison of the groups. For spiking and recovery studies, RSD (or CV) effect were determined accordingly,

$$\mathbf{E\_{CV}} = \mathbf{G\_{CV}} - \mathbf{M\_{CV}} \tag{6}$$

where ECV = CV effect, GCV = Group CV and MCV = Grand mean CV

The extraction methods ranked according to CV effect values. HorRat values and ORSD/PRSD were used in replication studies to rank sample splitting procedures. For repeatability data, PRSD values were calculated from modified Horwitz equation of PRSD < 2(1−0.5logC) <sup>×</sup> 0.67, while, for intermediate precision (within-laboratory reproducibility), PRSD was calculated from the unmodified Horwitz equation: PRSD < 2(1−0.5logC), where C is the concentration of the analyte. The intermediate precision data (from replication studies) were further subjected to one-way ANOVA to determine whether the day of analysis and the analyst affected the data. For LOD and LOQ data, outlier OD values were identified using the unmodified Horwitz equation: RSD < 2(1−0.5logC) where, C is the aflatoxin concentration. OD values above B<sup>0</sup> were also treated as outliers. The HorRat value associated with efficiency of extraction solvent of different shelf ages was used to evaluate variability due to solvent storage time.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2072-665 1/13/3/216/s1, Dataset S1. Aflatoxin recovery data, Dataset S2. Precision data, Dataset S3. LOD and LOQ data, Dataset S4. Robustness-ruggedness data.

**Author Contributions:** Conceptualization, J.K., T.W., R.M., J.F.L. and D.G.; Data curation, J.K.; Formal Analysis, J.K., J.F.L. and R.M.; Funding acquisition, D.G., J.F.L., J.K. and L.M.; Investigation, J.K., R.M., T.W., D.M. and L.M.; Methodology, J.K., R.M., T.W. and T.P.H.; Project administration, J.K.; Resources, J.K., T.W. and T.P.H., Supervision, J.K. and R.M.; Visualization, J.K.; Writing—original draft, J.K.; and Writing—review and editing, T.W., J.F.L., L.M., T.P.H., R.M., D.G., D.M. and J.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Consultative Group on International Agricultural Research (CGIAR)-Research Program (https://www.cgiar.org/research; accessed on 12 March 2021), Agriculture for Nutrition and Health (CRA between ILRI and KALRO Ref. Contract No. 1/2014)

and Government of Kenya through Kenya Agricultural & Livestock Research Organization (http: //www.kalro.org; accessed on 12 March 2021). The APC was funded by CGIAR.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available as supplementary material (Dataset S1. Aflatoxin recovery data; Dataset S2. Precision data; Dataset S3. LOD and LOQ data; Dataset S4. Robustness-ruggedness data).

**Acknowledgments:** This study was funded by CGIAR Research Program, Agriculture for Nutrition and Health (CRA between ILRI and KALRO Ref. Contract No. 1/2014) and the Government of Kenya (through KALRO) whom we sincerely thank. Former Director, Biotechnology Research Institute (BioRI) Sylvance Okoth is appreciated for his support and permission to carry out this work at KALRO-BioRI, Muguga. We acknowledge Andrew Slate (posthumously) of North Carolina State University, and Aloo of Public Health Laboratory(Kenyatta Hospital Complex, Nairobi) for their technical assistance. KALRO staff (Ben Wanyonyi, George Maina, Andrew Mageto, Gilbert Ouma, Sarah Kairuthi, Jackline Kagendo, Jacqueline Arusei, Agnes Nekesa, Mercyline Ong'ale and Clarah Jebet) also provided technical expertise while David Kinoti and M/s Joanna Auma (KALRO) provided statistical and bibliography softwares respectively. Phyllis Alusi of illustration unit KALRO, formatted the graphical abstract of this article. David Gikungu, the Deputy Director, Climate Services of the Kenya Metrological Department provided room temperature data for laboratory 5 (BioRI) for the period when this work was carried out. A former science teacher, Robert S. I. Karuku is highly acknowledged posthumously for inspiring the first author to the world of food poisoning, the main drive in this communication.

**Conflicts of Interest:** The authors declare no conflict of interest. TH is an employee of the manufacturer of aflatoxin diagnostic kit used in this study.
