Grain quality through the looking glass

A few one-minute infrared tests on a tiny three-gram sample from a few Durum wheat kernels are safeguarding the future of pasta quality while digital imaging shows potential as a realistic alternative to subjective quality control by eye. Ancient varieties can hold new possibilities for more sustainable agriculture in meeting nutritional needs.

Apart from the total solar eclipse on 20 March, there was little to distract from a packed program of interesting topics at this year’s Grain Network meeting in Milan. In keeping with the location, a number of experts from the Italian grain industry looked at the work going on to ensure quality durum wheat for pasta production through routine analysis at grain receival and in grain breeding programs. Proficiency testing, digital imaging assessment of grain quality and developments in NIR analysis provided further perspectives on the potential for improved global grain production.

Protecting pasta quality

Boiling water and a pinch of salt - pasta may be easy to cook, but behind the scenes, industry professionals need to maintain a positive relationship between the quality of durum wheat and consumers’ demands for quality pasta.

One particularly important quality parameter is the protein content of durum wheat. It must be at least 12%, but with modern farming under pressure this cannot be taken for granted. Dr Laura Gazza, Council for Agricultural Research and Economics - Research Unit for Cereal Quality, Italy said: “In Durum wheat production in Italy, protein content is decreasing year by year and this is a very big problem.”

The cereal quality network run by the council helps to monitor the situation by sharing data in real time and by promoting appropriate agricultural practices. “We try to make a connection between the grain quality data that we obtain by the NIR Infratec network and the agronomic practices throughout Italy,” she said.

Italy is the world’s second largest producer of Durum wheat, producing four million tons in 2014 against the six and a half million tons produced by Canada in the same year.

Dr Carla Corticelli, director of the Unione Seminativi national grain organisation, outlined the quality traits that any pasta producer wants from the grain. These include elasticity, colour and texture and the aforementioned protein content. Routine analytical tools like NIR instruments are important for speed and convenience. “They must give rapid, highly accurate results, under operating conditions that are not always easy. These tools are very important for breeding, not only for raw material, but because they allow you to use small quantities of material, which for breeding is very important,” she said.

A century of wheat research and development

“The importance of breeding programs was explained by Dr Marilena Paolini, PSB (Syngenta), Italy. PSB has been working with wheat breeding for over a century in Italy, the past 30 years of which has involved NIR. “For us it is important to do a lot of analysis in a very short time,” said Paolini.

NIR analysis helps to make it affordable by providing high throughput in the Syngenta PSB laboratory where around 200 ground and 500 unground grain samples are often tested in a single day. Small samples as low as three grams can be tested to preserve the sample for growing. The tests give a good and robust prediction of protein and hardness/colour, making quality screening in the early generation possible.

Ancient varieties still valuable today

According to consultant Dr Oriana Porfiri, breeding programs have an ocean of varieties and subvarieties of common wheat, minor cereals and ancient varieties to draw on.

Improved data about the characteristics of these grains is allowing researchers to see differences within species, finding, for example, new strains that are suitable for production on hilly or marginal land or with niche characteristics such as fibre or antioxidants. “By cultivation we can conserve the diversity, we can make use of marginal lands and we can answer growing demands of organic farming and organic products. Also in the new European policy, rural development is very important and these species can be a part of this policy,” concluded Porfiri.

Continued progress with digital imaging of grain quality

Whatever the grain variety, new imaging technology is making it easier to get a more consistent quality assessment in the form of the EyeFoss. The concept of using digital imaging to replace subjective assessment of grain for defects is quite revolutionary, replacing a subjective system that has been used ever since man started to grow grain. The first commercial use under a pilot project by CBH in Australia was a theme at last year’s meeting as presented by Ian Sproul, applied technology manager, CBH. This year, Ian returned to give an update on progress and use in a second harvest in Australia.

The work to get to this stage of commercial use has been considerable with close cooperation between CBH and FOSS essential to success. “For people who are not sure why we are spending nine years of our lives trying to develop this instrument, it is really to simplify the process of grading wheat and barley,” said Sproul. “The vision is to simplify the process so that we don’t need to train staff and we don’t need to have long discussions with our customers about things like grain damage.”

Progress has been made in improving the instrument, and interest from the industry is growing. Calibrations have been significantly improved, but there is still room for further improvement, drawing on more data from the grain industry about grain quality. Further implementation is planned. In conclusion, Sproul said: “It is no longer about getting the instrument to work, but more about rolling it out across the industry.”

Faster NIR and no more worries about standardisation

In comparison to digital imaging, NIR has been established for decades in the grain world, but much is still being done to develop this powerful analytical technology. FOSS product specialists brought attendees up to date with new developments in NIR, namely, a way of speeding up the analysis time for normal samples and higher levels of instrument standardisation. “We have done something quite simple with the Infratec NOVA,” said chemometrician Robin Malm.

He explained how the instrument makes a measurement based on typically 10 subsamples to get a representative result. The average of the sub-sample predictions is used to get the result. “But we also get the standard deviation,” said Malm. “If there is a big difference between subsamples, we get a high standard deviation and the other way round, if the difference between subsamples is small, the standard deviation is small.”

This principle can be used to speed up the measurement of normal samples as product specialist Bengt Nordlund explained. “If the standard deviation between the first five subsamples is low, it will only run those five,” he said. “If it is high, it will run 10 subsamples as it usually does.”

Nordlund also explained how the built-in wavelength calibration with the Infratec NOVA makes the instrument stable over a long time. “It will not drift so it takes away the need for yearly standardisation, but also the standardisation between instruments because all instruments are identical,” he said.

Stable measurements year after year

In addition to instrument stability, the calibrations used must also be stable to ensure reliable representative measurements year after year. The annual Infratec ring test gives a clear picture of this in relation to the FOSS ANN grain calibration. Presenting the results, chemometrician Tomas Nilsson said: “This international ring test is essential to validate against ISO standards, to avoid drifting with the calibration model and to monitor reference methods, local prediction models, the instruments and the reproducibility of the FOSS ANN prediction models.”

As in previous years, the FOSS ANN calibration shows great stability for protein and moisture in wheat and barley against reference methods. Good performance with the rapeseed prediction models was also seen, especially in terms of reproducibility, which is slightly better than the reference method.

Image credit: ©iStockphoto.com/Dimitrije Tanaskovic