Traditional vs. Modern Seed Identification Methods: An Overview

The iteration of seed variety identification technology reflects the transformation track of agriculture from experience to precision. Traditional methods rely on the naked eye to observe phenotypic characteristics such as seed morphology and field planting performance, and with the advantages of low cost and easy operation, they were once the main means of variety identification in agricultural production. Farmers distinguish varieties by intuitive differences such as grain size and plant type, while breeders ensure purity by field planting identification for several years. However, this method is easily disturbed by the environment, and it is difficult to deal with variety identification with a similar genetic background, with limited efficiency and accuracy.

The breakthrough of modern technology has achieved a qualitative leap: DNA labeling technology can capture the subtle differences at the gene level, and spectral imaging and artificial intelligence compress the detection period from several months to minutes, which fundamentally overcomes the limitations of traditional methods. However, the balance between the practical value of traditional methods and the application threshold of modern technology is still reshaping the technical path of seed identification. This collision not only continues the practical wisdom of agricultural production, but also promotes the upgrading of the seed quality control system.

The article provides an overview of traditional and modern seed identification methods, comparing their principles, advantages, limitations and applications.

Need for Seed Authentication and Purity Identification

As the basis of agricultural production, the authenticity and purity of seeds directly determine the yield, quality, and stress resistance of crops. Against the background of global food security facing multiple challenges such as climate change and population growth, accurate seed variety identification has become a key link to ensure the stability of agricultural production. Specifically, the importance of seed identification is reflected in three core dimensions:

• First, ensure crop yield and quality. There are significant differences in growth characteristics, yield potential, and quality indexes among different varieties of the same crop. For example, high-quality rice varieties need to have specific amylose content and cooking quality, and every 1% decrease in the purity of hybrid corn varieties may lead to a 2%~3% decrease in yield. If the planted varieties are inconsistent with the claims or the purity is insufficient, it will lead to inconsistent crop growth and decreased ability to resist pests and diseases, which will eventually lead to reduced production or even no harvest.
• Second, protect intellectual property rights and market order. Modern breeding has a huge investment, and the cultivation of a new variety often takes more than 10 years and hundreds of millions of yuan. Seed variety identification is the technical basis for safeguarding breeders' rights and interests, which can effectively curb illegal activities such as variety infringement. According to the statistics of the International Union for the Protection of New Plant Varieties (UPOV), the annual loss caused by seed counterfeiting in the world exceeds 10 billion US dollars, and the standardized identification system can provide a scientific basis for disputes over variety rights.
• Finally, support the global food security strategy. In cross-border seed trade, variety authenticity is the core index of quarantine inspection, which can prevent the spread of fake and inferior seeds carrying quarantine pests. At the same time, accurate identification is helpful to popularize excellent varieties adapted to local ecological conditions and enhance the anti-risk ability of regional agricultural production.

The average spectra with the standard deviation of five crop seeds (Wu et al., 2021)

From the perspective of technological development, seed identification methods have experienced the evolution from intuition to accuracy and from experience to science. Before the 20th century, farmers mainly relied on naked eye observation of seed morphology and planting performance to distinguish varieties; In the mid-20th century, biochemical technologies such as protein electrophoresis began to be applied.

Since the 21st century, the breakthrough of molecular markers and genome technology has made the identification accuracy reach the level of single base. Nowadays, traditional methods and modern technology are forming a complementary pattern and jointly constructing a multi-level seed quality control system.

Traditional Seed Identification Methods

The traditional identification method is based on the phenotypic and biochemical characteristics of seeds and plants, which has been used for a long time in agricultural production. Its core advantages are simple operation and low cost, and it is suitable for grass-roots production scenarios.

Morphological Analysis

Morphological analysis is the oldest and most widely used method to identify varieties by observing the appearance characteristics of seeds and plants. At the seed level, the identification indicators include:

• Physical morphology: Such as the grain shape of wheat seed (oblong/ovoid), the endosperm type of corn seed (hard grain/horse tooth), the color of rape seed (yellow/black), etc.
• Microscopic characteristics: Observe the seed coat texture and embryo morphology through a magnifying glass or microscope. For example, the difference between indica and japonica rice subspecies can be judged by the density of glume hairs.
• Physiological and biochemical characteristics: Identification is made by physiological reaction or chemical substance difference during seed germination.

Accuracies of each combination (Zhu et al., 2019)

This method has high practicality when the variety difference is significant, for example, ordinary wheat and black wheat can be quickly distinguished by grain color. However, the limitations are also obvious: on the one hand, the morphological characteristics are easily affected by the environment, and the same variety may show phenotypic differences under different planting conditions; On the other hand, for varieties with similar genetic background (such as sister hybrids), it is difficult to distinguish them accurately only by morphology, and the misjudgment rate can reach more than 15%.

Planting Identification Method

Grow-Out Tests (GOT) is a standard method recommended by the International Association for Seed Inspection (ISTA), which judges the authenticity and purity of varieties by planting seeds in the field and observing the phenotypes during the whole growth period. The operation process includes:

• Sample design: According to ISTA regulations, at least 400 plants (small crops such as rice and wheat) or 200 plants (large crops such as corn and cotton) should be planted; ​
• Characteristic observation: Record the key agronomic traits such as plant height, leaf shape, flowering period, and fruit shape, especially pay attention to the unique "fingerprint traits" of varieties (such as the red ear of sorghum and the inclination of sunflower).
• Purity calculation: the proportion of plants that do not meet the typical characteristics of varieties (that is, "heteromorphic plants") is the impurity rate, which is used to calculate the purity.

The advantage of GOT is that the results are intuitive and reliable, which can reflect the performance of varieties under actual growth conditions, so it has been included in the legal procedures of seed certification in many countries. However, its biggest drawback is that the cycle is too long, and it needs to wait for the whole growth period of crops (for example, it takes 3~4 months for rice and 5~6 months for wheat), which can not meet the needs of rapid detection. In addition, field experiments are greatly influenced by environmental factors such as climate and soil, and require a lot of land and labor costs.

Chemistry and Microscopy Technology

The chemical analysis method developed in the mid-20th century realizes variety identification by detecting the difference of specific components in seeds, which mainly includes:

• Protein electrophoresis: Identification was made by using the difference in electrophoresis patterns of different seed storage proteins (such as zein and wheat gluten). For example, soybean varieties can be divided into seven main protein types by polyacrylamide gel electrophoresis (PAGE) with a resolution of over 80%.
• Isozyme analysis: It is widely used in potato, rape, and other crops to distinguish varieties by detecting the zymogram differences of peroxidase, esterase, and other isoenzymes.
• Dyeing techniques: such as identifying the hardness of wheat seeds by phenol dyeing, or observing the difference in seed coat permeability by fluorescence dyeing.

The stability of these methods is better than morphological analysis, but there are still limitations: the expression of protein and isozyme is affected by the seed development stage, and for varieties with high genetic similarity (such as those bred by continuous backcross), the difference of biochemical markers is not significant, and the identification efficiency is greatly reduced.

Schematic representation of three approaches used for constructing spectral image cubes (ElMasry et al., 2019)

Common Limitations of Traditional Methods

Generally speaking, there are three bottlenecks in the traditional appraisal method:

• First, it is subjective and depends on the experience of appraisers, and different operators may get different results
• Second, environmental sensitivity, phenotypic and biochemical characteristics are easily disturbed by light, temperature, water and fertilizer conditions
• Third, the timeliness is poor, especially the GOT method, which can not meet the needs of rapid customs clearance and market supervision in seed trade

These defects promote the application of modern molecular technology in seed identification.

Modern Approaches for Seed Identification

With the development of molecular biology and detection technology, seed identification has gone from the phenotypic level to the genetic level, forming a modern system with DNA analysis as the core, combining spectrum, artificial intelligence, and other technologies, and its accuracy and efficiency have made a qualitative leap.

Molecular Marker Technology Based on DNA

DNA marker technology is the most accurate identification method at present, which directly detects the genetic material differences among varieties and is not affected by the environment. Mainstream technologies include:

• PCR and derivative technology: Polymerase Chain Reaction (PCR) can detect trace samples by amplifying specific DNA fragments. Among them, random amplification of polymorphic DNA(RAPD) and amplification of genomic DNA with random primers are easy to operate but have poor repeatability. Amplified fragment length polymorphism (AFLP) is a combination of restriction enzyme digestion and PCR technology with high polymorphism, which is suitable for crops with complex genetic backgrounds (such as sugarcane and potato). ​
• SSR marker: Simple sequence repeat (SSR), also known as a microsatellite marker, consists of 1-6 base repeats. Polymorphism is formed by the difference in repetition times of different varieties, which can be separated and detected by electrophoresis after amplification by specific primers. SSR markers have the characteristics of co-dominance and good repeatability, and have become the standard method for the identification of main crop varieties such as rice and corn.

Schematic representation of adapter-primers and UM-PCR (Wen et al., 2012)

• SNP marker: Single nucleotide polymorphism (SNP) refers to the variation of a single base in the genome (A/T/C/G), which is densely distributed in the genome (there is one SNP for every 100~300 bases). Large-scale SNP detection can be realized by gene chip or high-throughput sequencing technology, and hundreds of thousands of SNP loci can be analyzed by a single chip at a time. The advantage of the SNP marker lies in its scalability and automation, and it is suitable for batch sample detection. At present, SNP markers have been applied to the variety identification of wheat, soybean, and other crops, with an accuracy rate of 99.9%.
• DNA barcode technology: A conserved "standard sequence" (such as the chloroplast rbcL gene) in the genome is amplified and sequenced to establish a sequence database of species or varieties. When identifying, we only need to compare the sample sequence to be tested with the database, and we can quickly determine the variety ownership. This technology is simple to operate, especially suitable for rapid quarantine inspection in the seed trade.

Spectroscopy and Imaging Technology

Spectroscopy and imaging technology can realize non-destructive identification by analyzing the physical and optical characteristics of seeds, avoiding complicated operations such as DNA extraction, and are suitable for on-site rapid detection. ​

• NIR technology: Based on the difference of absorption of near-infrared light by organic components (protein, fat, and water) in seeds, a variety characteristic spectral model was established. Through partial least squares regression (PLSR) and other algorithms, wheat, rape, and other crop varieties can be quickly distinguished with an accuracy rate of over 90%. NIR has the advantages of fast detection speed (only 30 seconds for a single sample) and no damage, and has been included in the seed quality testing standards of some countries.
• Hyperspectral imaging technology: Combining spectral analysis and image recognition to obtain spectral information and morphological characteristics of seeds at the same time. For example, through hyperspectral images, the moldy areas of corn seeds can be identified, and the grain characteristics of different hybrids can be distinguished at the same time. This technology is outstanding in seed purity detection, which can automatically count heterogeneous seeds, and its efficiency is more than 50 times that of manual detection.
• X-ray imaging: By analyzing the internal structure of seeds (such as endosperm integrity, empty shrinkage rate), it assists variety identification, and is especially suitable for detecting hidden defects in seeds, such as rice chalkiness difference.

Automation and Artificial Intelligence

The integration of artificial intelligence technology greatly improves the efficiency and intelligence level of seed identification.

• Machine vision: The image of seeds is collected by a camera, and the characteristics of seeds, such as shape, color, and texture, are identified by a deep learning algorithm (such as a convolutional neural network, CNN). In the purity detection of soybean and vegetable seeds, the accuracy of machine vision can reach 95%, far exceeding 80% of manual identification, and it can work continuously for 24 hours.
• Multi-technology fusion model: integrating spectral data, image features, and DNA labeling results to establish a multi-dimensional identification model.
• Blockchain technology: The variety identification data (such as SNP fingerprints and spectral features) are stored on the chain to form an unbreakable "digital ID card". The whole process information of seeds from breeding to sales can be traced back, effectively curbing the counterfeiting behavior of varieties.

Relationship between model verification accuracy and iteration rounds (Bi et al., 2022)

Core Advantages of Modern Methods

Compared with traditional methods, the advantages of modern technology are concentrated in four aspects:

• First, accuracy, molecular markers directly reflect genetic differences and are not disturbed by the environment
• Second, speed, SNP chip and NIR technology can realize rapid detection of batch samples
• Thirdly, automation, artificial intelligence and robot technology reduce manual operation and human error
• Fourthly, the expansibility. Modern technology can continuously improve performance by upgrading algorithms and equipment to meet the new variety identification requirements

Comparative Analysis: Traditional vs. Modern Methods

There are significant differences between traditional methods and modern methods in technical principles, applicable scenarios, and performance indicators. Systematic comparison can provide a basis for practical application.

Accuracy

The accuracy of traditional methods is restricted by many factors: morphological analysis depends on the experience of appraisers, and it is easy to misjudge varieties with similar phenotypes, and the error rate of naked eye identification can reach 20%. Although the GOT method is intuitive, the difference of field environment may lead to phenotypic variation, and the calculation error of purity is about 5%.

The accuracy of modern molecular methods has been greatly improved: the accuracy of SSR markers in rice variety identification is over 98%, and SNP markers are 99.9%, which can distinguish varieties with genetic similarity as high as 99%. Spectral technology combined with AI algorithm has an accuracy of 92%~96% in wheat variety identification, and its stability is not affected by the environment.

Speed

Time cost is the significant difference between the two methods: GOT method needs to wait for crops to mature, and the cycle is as long as 3~6 months, which can not meet the rapid circulation demand of seed market; The detection period of morphological and biochemical methods is 1~7 days, but it is difficult to process in batches due to the limitation of sample size.

The detection speed of modern methods has achieved a qualitative leap: NIR and machine vision technology only take 1~5 minutes to detect a single sample, and can handle tens of thousands of seeds every day; SNP chip can detect 96 samples at a time through the automation platform, and it takes only 24 hours from DNA extraction to result analysis. Portable PCR equipment can even achieve rapid identification within 30 minutes in the field.

Accessibility

The advantage of traditional methods lies in their field applicability: farmers and grass-roots agricultural technicians can master morphological identification skills through simple training and complete preliminary screening in the fields. Although the GOT method is time-consuming, it does not need complicated equipment and is suitable for areas lacking laboratory conditions.

Modern methods rely on professional facilities and technology: molecular markers need laboratories and professional technicians, and spectral equipment needs to be calibrated regularly, which limits its application in remote areas, but the development of portable equipment is breaking this barrier.

On the whole, traditional methods are suitable for low-precision and low-cost preliminary screening, while modern methods are suitable for high-precision and large-scale detection requirements. In practical application, the two methods often complement each other: first, preliminary screening with morphology or NIR, and then molecular marker confirmation of suspicious samples, which not only ensures efficiency but also controls cost.

Comparison between Traditional and Modern seed identification methods

Conclusion

To sum up, traditional identification methods and modern identification techniques have irreplaceable values in the detection of seed variety authenticity and purity. The former focuses on morphological observation and field planting identification. Although it is restricted by environmental factors and phenotypic plasticity, its simplicity of operation and ecological authenticity are still the basic means of grassroots screening. The latter, with the help of molecular markers, spectral imaging, and other technologies, has realized the accurate analysis of genetic essence, significantly improved the detection efficiency and resolution, but faced the challenge of equipment cost and standardization system.

Future research needs to focus on the synergistic integration of the two, build a hierarchical detection system of "phenotypic screening-molecular confirmation", and promote technical standardization and portable research and development to meet the detection needs in different scenarios, and provide more systematic technical support for seed quality control.

References

1. Wu N, Liu F, Meng F, Li M, Zhang C, He Y. "Rapid and Accurate Varieties Classification of Different Crop Seeds Under Sample-Limited Condition Based on Hyperspectral Imaging and Deep Transfer Learning." Front Bioeng Biotechnol. 2021 9: 696292.
2. Zhu S, Chao M, Zhang J, Xu X, Song P, Zhang J, Huang Z. "Identification of Soybean Seed Varieties Based on Hyperspectral Imaging Technology." Sensors. 2019 19(23): 5225.
3. ElMasry G, Mandour N, Al-Rejaie S, Belin E, Rousseau D. "Recent Applications of Multispectral Imaging in Seed Phenotyping and Quality Monitoring-An Overview." Sensors. 2019 19(5): 1090.
4. Wen D, Zhang C. "Universal Multiplex PCR: a novel method of simultaneous amplification of multiple DNA fragments." Plant Methods. 2012 8(1): 32.
5. Bi C, Hu N, Zou Y, Zhang S, Xu S, Yu H. "Development of Deep Learning Methodology for Maize Seed Variety Recognition Based on Improved Swin Transformer." Agronomy. 2022 12(8): 1843.