Powered by FOSS

Accurate Nutrition Without the Lab

Technology Innovation Award Winner Adisseo takes the time and guess work out of nutrition.

Feed & Grain magazine have partnered with AgGateway and AFIA to develop the Information Technology Innovation Award. In this article re-published from Feed & Grain, editorial assistant Steven Kilger explains the reasoning behind this award and why Adisseo was chosen as this year’s winner for their web-based platform Precise Nutrition Evaluation (PNE).

Article by Steven Kilger
As published in the August/September 2013 issue of Feed & Grain

The United Nations’ department of Economic and Social Affairs released a report in 2012 that said the world’s population will reach an estimated 9 billion by the year 2050. This has placed a great deal of emphasis on agriculture and nonagriculture entities to determine the best ways to produce food for future generations. The current facts are that the amount of arable land to plow and sow is limited, and meat consumption is increasing as world economies evolve. Given the finite resources that are available for food production, technology will need to provide solutions for feeding an ever growing population. This is why Feed & Grain has partnered with AgGateway and the American Feed Industry Association (AFIA) to develop the Information Technology Innovation Award, acknowledging the work being done to make the process of putting food on the table more efficient and economical.

This year’s winner is Adisseo for its web-based platform, Precise Nutrition Evaluation (PNE). PNE combines an internet-accessible database with near infrared spectroscopy (NIR), in vivo (animal feeding studies) analyses of ingredients for the determination of apparent metabolizable energy (AME) and digestibility of amino acids (DAA), and ISO-certified wet chemistry methods that are both stringent and high quality. Together, PNE delivers an accurate, near-instantaneous nutritional analysis for the three most expensive nutrients within 25 feed ingredients.

Adisseo’s PNE service gives nutritionists, livestock producers and feed manufacturers more control over the nutritional composition of their feeds. In addition, purchasing decisions can be adjusted based upon ingredient quality, thus providing a better, more economical match between nutrition vs. animal requirements for higher feed quality and more efficient food production. Rob Shirley, poultry technical manager, and Steve Gately, director of rovabio enzyme business, of Adisseo help explain what PNE is and how it benefits livestock.

The problem

Prior to NIR, resources for determining the nutritional quality of feedstuffs were fairly limited, and for the most part, based on word-of-mouth, tables and wet chemistry. While suppliers, colleagues and friends provide word-of-mouth information, books and magazines provide a reference of “table values” for the nutritional content of ingredients. Both sources are capable of providing guidance on the nutritional composition of ingredients; however, the data from each are for all intents and purposes inaccurate because they represent a snapshot in time, can be extremely dated, do not reflect continuously changing circumstances that lead to nutritional variation (i.e., type of cultivar, geographic region, weather, harvesting, storage, manufacturing process, etc.), and can be from a mixture of wet chemistry procedures that may be poorly defined and/or conflict with each other. To get around this uncertainty and help ensure a certain level of animal performance, ample safety margins for key nutrient matrices are often built into formulations with the idea that it is better to have too much than too little. This over-formulation strategy often costs money, and may not necessarily result in better animal performance.

In order to reduce safety margins, feed manufacturers and integrators can analyze the nutritional content of ingredients on a regular basis using wet chemistry and in vivo analyses. The information that is generated is considered more dynamic and “real time” because it not only reflects analytical information that is generated on a regular schedule, but is also sometimes derived from actual in vivo (in life) feeding studies. As data are made available, they can be entered into a database to show trends over time. Unfortunately, in any quality control program, analytical reports are often received days to weeks after the ingredients of interest are consumed by livestock. Putting this into perspective, Gately explained that “sending a sample out for metabolizable energy and digestible amino acids tests in the past would take you weeks and weeks to get an analysis back.”

In order to turn these dynamic data sets into a more useful tool that report historical data in real time, NIR has been adopted by many feed manufacturers and livestock integrators because it is rapid, reliable, accurate, nondestructive and relatively inexpensive. Today, in-house NIR equipment and PNE have cut the nutrient analysis time of ingredients down to mere minutes.

The PNE proposition

Using NIR to determine the proximate analyses (protein, fat, fiber, ash and moisture) for individual ingredients, and to a lesser extent mixed feeds, has been around for years. So when Adisseo started to work with NIR technology over 10 years ago, they knew they wanted to do something different and innovative. This is when PNE was conceived and developed.

“At Adisseo, we made the decision to take NIR one step further and develop DAA and AME calibrations. To do this, Adisseo had to put an extensive amount of resources into developing NIR calibrations and the PNE database,” Gately noted. “As a result, Adisseo started testing ingredients in vivo at their research facility in Commentry, France, using the rooster assay for DAA and the 3-week-old broiler assay for AME analyses. In a typical AME trial, 3-week-old chicks are fed a complete diet and a diet where part of it is replaced with the test ingredient. The excreta are then collected, freeze dried and analyzed for moisture, gross energy and nitrogen content. The comparison between the energy in the two diets vs. the respective excreta is made to determine the amount of energy (AME) that the bird was able to extract from the ingredient. The AME trial setup for each ingredient is replicated several times for statistical purposes, and is therefore a very laborious, time-extensive and expensive possess.”

This process resulted in AME calibrations for corn, wheat and soybean meal, along with total and phytate phosphorus information for nine ingredients and total and digestible amino acid calibration equations for 25 ingredient types.

The process

The first step in obtaining an accurate analysis is grinding. Grinding a sample ingredient to 1 millimeter (this is the same as 1,000 microns) reduces particle size and increases the uniformity of particle distribution throughout the sample that is scanned on the NIR. Grinding also allows for a more complete penetration, reflection and detection of light from the sample. It is important that samples are not ground too fast, as this will overheat and change the nutrient composition of the sample. When the customer uploads a spectral file to PNE via internet, the user defines the sample type, origin of sample and the analyses desired. Once submitted, PNE compares the sample spectra to its core calibrations for the chosen ingredient and predicts the AME, total and phytate phosphorus, and total and digestible amino acids within approximately two to three minutes.

“When we first made Adisseo’s NIR service available to its customers, pre-PNE, it took two to four hours to turn sample spectra around into actual data. Today, PNE’s calibrations are encrypted in the cloud, analyses are returned within two to three minutes, and data are always available to the customer in case they want to create trend charts for purchasing or nutrition-related activities,” said Shirley.

Near-instant turnaround time is unique to PNE, and has the ability to significantly change how feeds are formulated, ingredients are purchased and managed, and how feed cost is minimized.
After the drought of 2012, feed prices shot up to record-breaking prices and highlighted how flexibility can be an asset when formulating feed. Amino acids, energy and phosphorus are the most expensive components in a feed formula, and Adisseo’s PNE gives accurate data for all three at the same time. This allows nutritionists to change the blend of feed ingredients and formula specifications, knowing that the formula will meet the nutritional requirements for healthy and efficient growth and production.

In order to validate the usefulness of PNE in production, “Adisseo has conducted a number of broiler performance trials, where book values were compared against PNE values for AME and DAA. In all the trials, improvements in feed conversion were consistently observed when nutrient values from PNE were used. These trials clearly demonstrate the biological and financial value of knowing the nutritional composition of feed ingredients,” Gately said.

PNE gives buyers and purchasing groups the ability to verify that what they are receiving is what they are paying for, and also check where the best value can be found.

“If you look at the purchasing side of the equation, an integrator may have a number of suppliers to choose from. If a buyer is evaluating three distiller suppliers, the buyer has the option of ranking them according to how much digestible lysine is in each product. In addition, the amount of phosphorus and/or AME from different soybean meal suppliers can be a criterion by which to rank soybean meal suppliers. At this point, buyers can start to rank the value of each supplier’s product based on what they traditionally deliver,” Shirley remarked.

Precision is necessary in any science, and agriculture is a science that bases its productive output on cause-and-effect principles. In the case of swine and poultry production, knowing the nutritional makeup of feed ingredients is critical to formulating feeds that provide the correct nutrients, at the right levels. Nutritionists, livestock producers and feed manufacturers that use Adisseo’s PNE platform have more control over nutritional decisions for their livestock, create better business opportunities on ingredient selection and purchases, and provide a better match between nutrition and production in terms of higher feed quality, more efficient production and better profitability.

Want to know more about the potential and application of near infrared technology in feed production? Join our fast growing network of experts and professionals from both NIR and feed sectors now.

Christian Tollebaeck
Christian Tollebaeck

Kim, these are all very good and relevant questions and if I was a customer to one of these online calibration providers, no matter who it is, I would ask these questions directly to them as they should be able to answer.

Spectral outlier detection can be used such as Global H, Neighborhood H (some use Mahalanobis distance) and T-stats as the first indication if the sample is represented in the calibration, and that should be your first indication if the results is trustworthy. But of course I would also like to have more information about the database, reference methods and the calibration performance. 

I recommend everybody who is interested in this topic to obtain the ISO 12099 publication "Animal feeding stuffs, cereals and milled cereal products -- Guidelines for the application of near infrared spectrometry", I think it is a great guideline.

Kim Albert Perlado
Kim Albert Perlado


Thank you for sharing the ISO guideline. Potential users of these online provider should really look on the parameters you mentioned. I had a conversation with them and their service capabilities and their NIR model performance differs a lot. As in a wide difference!

ohh... And they also conduct ring test among their clients.. One sends monthly physical samples while others runs yearly and random test.

Ultimately, what you want to measure, the error you are willing to take, the sums of benefits it will bring and weight of money you can take out from your pocket are the deciding factors...

Kim Albert Perlado
Kim Albert Perlado

Normally you build a calibration or transfer the calibration file on another equipment and then perform a validation test. Validation will then tell you if you have a robust calibration or if your calibration transfer is successful. However in this platform you don't own the calibration; the sample spectra is sent to the web and the result is predicted online. What QC measures are used to ensure the result generated by the system is accurate and the instrument used by the client to generate the original spectra is in agreement with the NIR of the service provider?

Is it safe to assume that the online NIRS provider do have a rigorous QA procedure and as long as you can demonstrate that your NIR is performing well then the online results is accurate?