| Reason, D A, Watts, M J, and Devez, A. 2015. Quantification of phytic acid in grains. (Inorganic Geochemistry, Centre for Environmental Geochemistry). British Geological Survey Internal Report, OR/15/070.
|
A. Oat flour reference material
Although no certified reference material was commercially available, a sample of oat flour with established phytic acid values was provided with the phytic acid assay kit (Megazyme®) was initially used to measure method accuracy. The average measured value of 17 862 ± 838 mg kg-1 showed good agreement with the established concentration of 17 700 mg kg-1, representing a bias of 1% (Table 1).
Table 1 Accuracy and precision of measured values versus target values for an oat flour reference material (n=10)
|
|
Phytic acid
|
| Target Value (mg kg‐1)
|
17 700
|
| Measured mean (mg kg‐1)
|
17 862
|
| Standard Deviation (mg kg‐1)
|
838
|
| Precision (%RSD)
|
5
|
| Bias (%)
|
+1
|
B. Precision and accuracy for high and low standards
Standards at both 1790 and 23 679 mg kg-1 phytic acid (7 and 80% of the top calibration standard) were analysed (n=5). Since a phosphorus calibration standard (not phytic acid), was used, phosphorus concentrations were also included in the analysis. The 7 and 80% standards corresponded to 517 and 6000 mg kg-1 phosphorus, respectively. Tables 2 and 3 display the accuracy and precision for each standard. The accuracy of measurements for a solution equivalent to 7% of the top standard concentration was 92% for phosphorus and 94% for phytic acid. Table 3 shows the accuracy for phosphorus and phytic acid in a solution equivalent to 80% of the top standard concentration was 99 and 90% respectively.
Note: A 7% standard was used in place of a 5% standard due to the limits of quantification stated previously (a higher concentration standard was required to obtain reliable data).
Table 2 Accuracy and precision of a phytic acid measurement at 7% of the top calibration concentration (n=5)
|
|
Phosphorus
|
Phytic Acid
|
| Target value (mg kg‐1)
|
517
|
1790
|
| Mean measurement (mg kg‐1)
|
474
|
1679
|
| Bias (%)
|
‐8
|
‐6
|
| Measured SD (mg kg‐1)
|
34
|
122
|
| Precision (%RSD)
|
7.2
|
7.2
|
Table 3 Accuracy and precision of a phytic acid measurement at 80% of the top calibration concentration (n=5)
|
|
Phosphorus
|
Phytic Acid
|
| Target value (mg kg‐1)
|
6000
|
23679
|
| Mean measurement (mg kg‐1)
|
5970
|
21233
|
| Bias (%)
|
‐1
|
‐10
|
| Measured SD (mg kg‐1)
|
298
|
902
|
| Precision (%RSD)
|
5.0
|
4.2
|
C. Within and between run variation
To calculate both within and between run variation, four samples (2 rice and 2 wheat grain samples), were measured in triplicate (n=5*). Anova analysis was then used to calculate each source of variation (within and between run). All within run variations were below 10% and hence passed the validation criteria set in Appendix 2. Between run variation was often higher (with rice 1 as the exception). Rice 2 had extremely low levels of phytic acid verging on the LOQ for the analysis. To reduce between run variation in these low-phytate samples there are two possible solutions. More sample could be used in each assay (e.g. 2 gram of sample for the same extracting volume). This would raise the measured phytic acid concentrations (reducing the variation), which would be accounted for in the calculations. Alternatively all samples with measured phytic acid concentrations below 2000 mg kg-1 will be run in duplicate to ensure more accurate results. Large between run variations seen in the grain samples was due to a lack of homogeneity within the sample. As phytic acid is predominantly stored in grain husks, homogeneity has a significant influence on between run variations. Again this variation could be reduced in two ways. A larger sample and extracting acid volume (e.g. 3g sample, 60 mL acid), would allow for a more representative sample being analysed with no need for further sample preparation. Alternatively, additional sample preparation techniques could be investigated such as the use of a mortar and pestle to produce a finer more homogenised sample material. All results for within and between run variation can be found below (Table 4).
Note n=4 used for rice 1.
Table 4 Within and between run variation of an analysed rice
and grain sample using %RSD
|
|
Rice 1
|
Rice 2
|
Grain 1
|
Grain 2
|
| Between run (%RSD)
|
1
|
33
|
15
|
10
|
| Within Run (%RSD)
|
5
|
9
|
7
|
8
|
D. Overall precision
Overall precision was calculated based on all of the data obtained from two rice and two grain samples (Table 5). Rice 2 had very low phytic acid concentrations that approached the limit of quantification (1408 mg kg-1). Hence it is seen to have an associated high %RSD of 19%. Thus, any sample with a measured concentration below 2000 mg kg-1 should be run in duplicate to ensure accurate results. It was also noted that the wheat grain samples were less homogenous than the rice. Although the %RSD is still acceptable (≤10%), further sample preparation may result in more accurate results in future.
Table 5 Overall Precision of phytic acid for two rice and two grain samples (n=15)
|
|
Rice 1
|
Rice 2
|
Grain 1
|
Grain 2
|
| Overall Average (mg kg‐1)
|
6442
|
1534
|
10964
|
9809
|
| Overall SD (mg kg‐1)
|
271
|
292
|
1132
|
899
|
| Overall Precision (%RSD)
|
4
|
19
|
10
|
9
|
E. Spike recovery
A solution spiked with a calibration standard was used to determine the percentage recovery in a rice sample with a known concentration of phytic acid. 2 mL of a 100 000 mg kg-1 (assuming 1 kg L-1) solution was added to the extracted sample containing 20 mL HCl. This provided a total spike of 9090 mg kg-1, resulting in a total target concentration of 15 532 m kg-1. Table 6 displays an observed recovery of 94%. This value was calculated after the exclusion of an outlier using the Dixons test (Table 7). The likely reason for the outlier is the presence of a large air bubble during a pipetting step.
Table 6 Spike recovery data (n=4)
|
|
Phytic Acid
|
| Target spike value (mg kg‐1)
|
9090
|
| Mean spike measurement (mg kg‐1)
|
8550
|
| Average recovery
|
94%
|
| Measured SD (mg kg‐1)
|
65
|
| Precision (%RSD)
|
0.8
|
Table 7 Dixons test to exclude a result as an outlier
| Spike concentration (mg kg‐1)
|
Dixons parameter
|
Value
|
| 7084
|
Gap
|
1385
|
| 8470
|
Range
|
1531
|
| 8525
|
Q‐Score
|
0.905
|
| 8589
|
Q‐score for exclusion (99%)
|
0.821
|
| 8615
|
Result
|
Excluded at 99%
|
F. Analyst variation
A Students T-test was used to evaluate the null hypothesis; no significant variation occurred between multiple analysts. Grocery store bought red split lentils were analysed to confirm this hypothesis. First, an F-test confirmed equal variance of the two datasets with a value of 1.35, (lower than the critical value 8.85). The students T-test produced a T-value of 0.11 which was much lower than the 2.20 critical value. The null hypothesis was confirming at the 95% confidence level (Appendix 2) with a P value of 0.91 (Table 8).
Table 8 F-Test and T-Test results for analyst variation
|
|
F‐Test
|
T‐Test
|
| Calculated Value
|
1.35
|
0.11
|
| Critical Value
|
8.85
|
2.20
|
| P Value
|
0.44
|
0.91
|