|
|
|
© 2006 Plant Management Network. Improving Corn Grain Purity by Using Color-Sorting Technology A. Susana Goggi, Assistant Professor, Department of Agronomy, Iowa State University, 166 Seed Science Center, Ames 50011; Kamal M. Adam, Former Assistant Scientist II, Iowa State University, 155 Seed Science Center, Ames 50011; Higinio Sanchez, Graduate Student, Department of Agronomy, Iowa State University, 166 Seed Science Center, Ames 50011; and Mark Westgate, Professor, Department of Agronomy, Iowa State University, 1577 Agronomy Hall, Ames 50011 Corresponding author: A. Susana Goggi. susana@iastate.edu Goggi, A. S., Adam, K. M., Sanchez, H. L., and Westgate, M. 2006. Improving corn grain purity by using color-sorting technology. Online. Crop Management doi:10.1094/CM-2006-0309-01-RS. Abstract Color sorting is often used to remove unwanted off-colored contaminants. The objectives of this study were to determine the usefulness of color sorting in removing adventitious corn from a seed lot and to evaluate the impact of contaminant color on efficacy. Seed lots of two contrasting colors were used in the experiments: yellow corn in a white lot and purple corn in a yellow lot. Samples were collected from three experimental sites: two white corn seed production fields surrounding a yellow pollen source, and a yellow corn seed production field surrounding a purple popcorn pollen source. Collected samples were color sorted in three successive passes. Outcross levels in the original nonsorted samples ranged from 0.10 to 38.55% in the yellow and white samples and from 0.61 to 45.78% in the purple and yellow samples. Color sorting reduced the percentage of outcross of yellow and purple kernels in the sample. The percentage of yellow seeds in white-seeded corn was reduced to 6.22% in samples collected at 3 ft and 0.01% at 820 ft from the source. The percentage of blue seeds in yellow-seeded corn samples was reduced to 0.58 at 0 ft and 0% at 270 ft from the source. These results indicate that color sorting is very effective at removing outcross seeds. Introduction For many years, the seed corn industry has produced seed of high genetic purity by effectively managing the reproductive stage of the parental lines. The use of high-purity parental seed, proper isolation distances (12), male border rows, and management of flowering synchrony and morphology (10) has allowed seed companies to consistently produce seed with varietal purity levels of 99% (4). Mechanical barriers, such as hedges or trees, and buffer crops are used by organic farmers to limit the flow of pollen from other pollen sources into grain and seed production fields (13). These control systems effectively achieve 99% or higher genetic purity of a product; however they fall short of the zero tolerance currently required for new plant-made industrial and pharmaceutical products (21). Precautionary measures, such as greater spatial and temporal isolation, are used to prevent unwanted outcross (1). But even these precautions may not prevent outcross under adverse circumstances, such as extreme climatic conditions leading to a breach in containment or human error (3). Adventitious seeds, those from another source, are likely to have similar physical characteristics as the desired product and thus are not easily removed using conventional conditioning equipment. Countries around the world have a variety of labeling and traceability requirements for plant-based biotechnology products, ranging from no regulatory or labeling requirements to elaborate and well-defined regulations that may or may not include mandatory labeling for biotechnology products (22). The European Commission (EC) (7) has established a tolerance of 0.5% adventitious presence of approved transgenic events in a nontransgenic product for feed or food. In seeds, the tolerance level is 0.3% for self-pollinated crops and 0.5% tolerance for cross-pollinated crops. However, the tolerance for events not approved by the EC is 0% (8,9). Adventitious seed can be removed from a seed lot if it differs in color. However, most of the current or pending plant-made pharmaceutical (PMP) or plant-made industrial (PMI) product transgenic corn varieties are yellow dent corn. Photo-optical detectors capable of distinguishing color differences were first developed in 1950s and evolved as an industry during the early 1980s (15). Today, color sorting is used in a variety of industries to remove unwanted off-colored contaminants. The food industry uses color sorting for the detection and removal of defective products from a wide range of cereals and snack foods, including baked or fire-extruded products, chips, wrapped candies, and dried fruits. Color sorting is used by the seed industry to detect Karnal bunt in wheat seed (6), to remove off-colored seeds (brown, tan, and green) from white-seeded snap beans to improve quality (14), and to segregate red from white wheat (17). No reports have been found in the literature where color sorting was applied to identifying and removing small amounts of outcross based on color contrast between product and contaminant. The objectives of this research were to evaluate the usefulness of an optical color-sorter for removing adventitious corn from a seed lot and to determine the efficiency of the color-sorter in removing the adventitious seeds. Grain Samples Grain samples used in this experiment were hand-harvested from three fields planted in Iowa in 2003. Two fields were located on ISU farms near Ankeny and one site was the ISU Allee organic-farm near Newell. These experiments were designed to assess outcross at different distances from a central pollen source field. We used these samples rather than mixing new grain samples because they had a broad range of adventitious off-colored corn kernels and presented a real case of contamination. The samples also provided two color contrasts: adventitious yellow corn kernels in white corn seed lot and adventitious blue corn kernels in yellow corn seed lot. Because the field experimental designs were developed to create outcross events, adventitious presence (2) of an off-colored seed in the corn samples was equivalent to an outcross event. Fields Where Samples of Yellow-Seeded Corn in White-Seeded Corn Were Obtained Table 1 summarizes field information. Samples with adventitious yellow corn were obtained from two fields of approximately 89 acres each planted with white corn hybrid RX792W. In the center of both fields, a 2.47-acre plot of DKC69-71 yellow corn was planted as an adventitious pollen source. Fields were managed under normal production practices of cultivation and insect and soil fertility management. Before flowering, one field was mechanically detasseled in a 4:1 ratio to reduce the local pollen density, as in a typical seed production field. The term "detasseling" in our study refers to mechanically removing the immature tassels (male inflorescence) to reduce the amount of pollen produced by the white corn. Thus, the chances of the incoming pollen from the yellow corn to "outcompete" the local white pollen increased and the number of outcrosses also increased. The two hybrids had synchronous flowering and the yellow corn pollen flowed freely into the white seed corn. Figure 1 shows a diagram of the field indicating the sampling locations. Because the level of outcross decreases with distance from the source field, sample size was increased at locations farther from the center of the field, according to calculations using Seedcalc6 (19). Seedcalc 6 is a Microsoft Excel (Redmond, WA) application that can be downloaded free from the International Seed Testing Association (ISTA) to help in the calculation of the number of seeds required for seed and grain purity testing. At harvest maturity, 25 ears were collected at 3, 33, and 115 ft from the adventitious yellow corn source, and 100 ears were collected at 328, 492, 656, and 820 ft from the adventitious yellow corn source along each of the four sampling transects (northeast, northwest, southeast, and southwest) (Fig. 1). Ears collected were dried to 12% moisture content and shelled with a laboratory-size sheller (Model LS 91, Custom Seed Equipment, Altoona, IA). The shelled corn was aspirated to remove all broken and light material using a laboratory size aspirator (Model 49470, Carter Day, Minneapolis, MN). Samples from both fields were combined into a group called "yellow-seeded corn outcross in white-seeded corn" and analyzed together.
Table 1. Fields planted in Iowa in 2003 where samples were collected to use in these color-sorting experiments.
x Plants were detasseled at a 4:1 ratio, four rows were detasseled and one row was not-detasseled. y N, NE, NW, S, SE, and SW = north, northeast, northwest, south, southeast, and southwest, respectively. Field Where Samples of Purple-Seeded Corn in Yellow-Seeded Corn Were Obtained Corn samples for the color-sorting experiments were obtained from an organic farming demonstration plot design to examine the distances purple-seeded popcorn pollen travels to a neighboring field. Figure 2 illustrates the field design and sampling locations. A strip of 24 rows of purple popcorn was planted within a 15-acre field of yellow corn. At harvest, a sample of 25 ears was collected at 0, 30, 60, 90, 120, 150, 180, 210, 240, and 270 ft from the adventitious purple popcorn source following three transects perpendicular to the adventitious pollen source, one on the south side and two on the north side. The samples were collected close to the source to demonstrate in the field the incidence of outcross in proximity in adjacent conventional and organic fields. Collected ears were processed in the same manner as described in the preceding section.
Color Sorting A 20-channel ScanMaster color sorter (Model SM-200DE, Satake USA, Stafford, TX) was used for color sorting. The machine consists of two closely spaced 10-channel chutes with a variable speed vibratory feeder. As the grain travels down the feed chute, the kernels accelerate and separate as they are guided to the machine’s viewer, which is illuminated by four incandescent lamp assemblies: two in the front and two in the rear. The optical assemblies measure the light reflected from each kernel of grain. When kernels of abnormal color are detected, they are diverted from the stream by a small blast of compressed air. These unacceptable kernels are collected in the reject chute. The acceptable kernels continue on their normal path into the accept chutes. An FSI-K filter set and a -3 with an 80% visible background were used with sensitivity of 348 and a feed rate of 330 (683 lb/h per chute) were used to separate the white and yellow corn mixture. An XEI-K filter set, a marigold -1 background, sensitivity of 750 and feed rate of 424 (972 lb/h per chute) were used for yellow and purple corn mixture sorting. Because all corn kernels in the samples were of similar size as determined by weight of 100 seeds (data not shown), delay and dwell switch settings of 13.7 and 12 corresponding to 12.9 and 0.95 milliseconds, respectively, were used to sort all samples. The color-sorter accuracy depends on the feeding rate but not on the size of the sample sorted. Sample size ranged from 10,000 to 70,000 kernels. The samples were weighed before and after each pass. Each sample was passed through the color-sorter three times, beginning with the original nonsorted sample and followed with the accept fraction of the previous pass. The accept fraction was weighed after each pass before passing through the color-sorter again. The reject fraction obtained after each pass also was weighed and recorded. The genetic purity of the initial sample, reject, and accept fractions was determined after each pass by manually separating the adventitious seed from the desired seed and weighing the fractions. Based on these results, the percentage of adventitious corn remaining in the accept fraction was then computed as the weight of the adventitious seed remaining in the sample × 100 × (total weight of the accept fraction)–1. Statistical Analysis The samples of yellow-seeded and white-seeded corn from both fields were pooled by distance (eight samples per distance). Similarly, the samples of purple-seeded and yellow-seeded corn also were pooled by distance (nine samples per distance). Three subsamples per sample were color sorted. A one-way analysis of variance (ANOVA) was conducted by distance with passes through the color-sorter as the main effect. All treatment effects were analyzed and results reported based on the accept fraction. The statistical package used was Statistical Analysis System (20). Means of successive passes were separated using Duncan multiple range test. Removal of Adventitious Yellow-Seeded Corn from White-Seeded Corn Adventitious yellow grain in the original nonsorted samples ranged from 38.55% at 3 ft to 0.10% at 820 ft from the central pollen source (Table 2). The percentage of adventitious yellow seed decreased significantly (P = 0.05) with the first color-sorting pass for corn samples harvested at 3, 33, 115, and 328 ft from the pollen source. Two passes through the color-sorter were necessary to significantly (P = 0.05) reduce the percentage of adventitious yellow seed from the unsorted corn samples harvested at 492 and 656 ft. No significant reduction was achieved by color-sorting corn samples harvested at the farthest distance (820 ft) (Table 2). No further reductions in adventitious presence occurred with successive color-sorting passes, except for the third pass with the seed collected at the 3- and 328-ft distances. In general, the efficacy of color-sorting in removing adventitious yellow seed was more readily evident in samples that originally had higher percentages of adventitious yellow seeds. The final percentages of adventitious yellow grain following three color-sorting passes ranged from 6.22% at 3 ft to 0.01% at 820 ft. Table 2. Mean percentage of adventitious yellow corn in the original samples and the accept fraction of color-sorted white corn samples collected at increasing distances from the yellow corn source field after three successive passes through a color-sorter.
Means within a column followed by the same letter are not significantly different at the 0.05 level of probability. x 0 = original nonsorted sample; 1 = after one pass through the color-sorter; 2 = after two passes through the color-sorter; 3 = after three passes through the color-sorter. Adventitious Purple-Seeded Corn in Yellow-Seeded Corn The initial adventitious purple seed in the samples ranged from 45.78% at 0 ft from the source field to 0.61% at 270 ft (Table 3). The percentage of adventitious purple seed significantly (P = 0.05) decreased with the first pass through the color-sorter for seed samples collected at all distances from the purple corn pollen source except at 270 ft (Table 3). Further reductions (P = 0.05) in purple seed occurred after the second pass on samples collected at 0, 30, and 240 ft from the purple source field. The percentage of purple seed did not decrease significantly after three passes through the color sorter. The final percentage of purple corn ranged from 0.58% in samples collected at 0 ft to 0% in samples collected at 240 and 270 ft from the purple corn pollen source. Conclusions Sorting by color effectively removed the adventitious seeds from the corn samples. The adventitious seeds were removed completely when the color contrast was pronounced and the initial percentage of adventitious seed was small. These findings agree with Pasikatan and Dowell (17) who found that separation of red from white wheat was more efficient than that of weather-discolored seed of the same variety. Similarly, the aflatoxin level in infected peanut seeds was successfully decreased using color sorting only when there was a marked color difference between the aflatoxin-containing and aflatoxin-free peanut seed (5). Generally, the levels of adventitious seed decreased by 10-fold after three passes through the color-sorter. When the seed lot’s initial outcross was approximately 40%, the lowest possible percentages achieved after three passes through the color-sorter was 6.22% for yellow adventitious seed in white seed corn lots and 0.58% for purple adventitious seed in yellow seed corn lots. However, the majority of the reduction was obtained in the first pass through the color-sorter. For example, to meet a 95.5% level of purity in a nontransgenic grain lot required by the EC, one pass through the color-sorter was necessary in a corn lot with approximately 1% adventitious presence if the grain had subtle color differences (yellow and white). The same level of purity (i.e. 95.5%) can be achieved by color-sorting a seed lot with an initial adventitious presence of 4.22%, if the adventitious seed had a greater color contrast (i.e., purple and yellow). For most seed lots, there was no statistical significance in increasing the number of passes beyond two passes. Most of the reduction in contamination was achieved after the first pass. The decision on the number of passes through the color-sorter should be based on the desired level of purity in the final lot. None of the containment methods currently used for open-field corn production are completely effective (11). Unless the production of PMP or PMI products is restricted to underground caves (16), it will be difficult to eliminate all risk of an escape in open-field production (18). Our results indicated that involuntary mixing can be removed if there is color contrast between the kernels. Unfortunately, most of the transgenic events currently approved for open-field production are on yellow corn. Color sorting is also a useful method for separating seed differing in color when human error or unusual climatic conditions produce an unwanted seed mix. Color coding of the product immediately after harvest could be an alternative to reduce involuntary mixes after harvest. However, when there is a slight color contrast between the lot and the contaminant, complete separation to achieve a tolerance level of zero is impossible. Acknowledgments We thank Dr. Ozkan Zengin (ISU Center for Survey Statistics and Methodology) for help with the statistical analysis of the data. This journal paper of the Iowa Agricultural and Home Economics Experiment Station, Ames, Iowa, Project No. 3638, was supported by the Hatch Act and the State of Iowa Funds, and by a USDA–CREES BRS Grant. Literature Cited 1. Baltazar, M. B., and Schoper, J. B. 2002. Crop-to-crop gene flow: Dispersal of transgenes in maize during field tests and commercialization. Pages 24-33 in: Proceeding of the Seventh International Symposium on the Biosafety of Bioengeenered Organisms, Beijing, China, October 10-16, 2002. 4. Burris, J. S., and Lauer, M. J. 2001. Adventitious pollen intrusion into hybrid maize production fields. Tech. rep. for the Assoc. of Official Certifying Agenc. 5. Chiou, R. Y. -Y., Wu, P. -Y., and Yen, Y. -H. 1994. Color sorting of lightly roasted and deskinned peanut kernels to diminish aflotoxin contamination in commercial lots. Am. Chem. Soc. 42:2156-2160. 6. Dowell, F. E., Wang, D., Baker, J. E., Throne, J. E., Steele, J. L., and Delwiche, S. R. 1997. Automated single wheat kernel quality measurement using near-infrared reflectance. Presented at the August 1997 ASAE Ann. Int. Meeting. Paper No. 97-6100. 7. European Commission (EC). 2004. Reg. No. 641/2004. Official J. Euro. Union. 7.4.2004. L102/16-25. 10. Fonseca, A. E., Westgate, M. E., Grass, L., and Dornbos, D. L. 2003. Tassel morphology as an indicator of potential pollen production in maize. Crop Management, doi:10.10.94/CM-2003-0804-01-RS. 12. Jones, M. D., and Brooks, J. S. 1950. Effectiveness of distance and border rows in preventing outcrossing in corn. Okla. Agric. Exp. Stn., Tech. Bull. No. T-38. Stillwater, OK. 13. Jones, M. D., and Brooks, J. S. 1952. Effect of tree barriers on outcrossing in corn. Okla. Agric. Exp. Stn., Tech. Bull. No. T-45. Stillwater, OK. 14. Lee, P. C., Paine, D. H., and Taylor, A. G. 1998. Detection and removal of off-colored bean seeds by color sorting. Seed Technol. 20:43-55. 15. McQueen, K. 2000. Optical sorting systems used for snack food quality control. Cereal Foods World 45:464-465. 17. Pasikatan, M. C., and Dowell, F. E. 2003. Evaluation of a high-speed color sorter for segregation of red and white wheat. Appl. Eng. Agric. 19:71-76. 18. Poppy, G. M. 2004. Geneflow from GM plants – towards a more quantitative risk assessment. Trends Biotech. 22:436-438. 19. Remund, K., Simpson, R., Laffont, J.-L., Wright, D., and Gregoire, S. 2005. Seedcalc6. Online. Statistical Tools for Seed Testing, Excel file, Int. Seed Testing Assoc. 20. SAS Institute. 1990. SAS/STAT user’s guide, version 6, 4th ed. SAS Institute, Inc., Cary, NC. 21. USDA–Animal and Plant Health Inspections Services. 2003. Field testing and plants engineered to produce pharmaceutical and industrial compounds. Fed. Regist. Vol. 68, No. 46. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
