Issue #98 | The weird link between food testing and the amazon rainforest | This SOP is hot! | Ultimate training decision tool) | Watermelon Chart |

2023-07-24

This is The Rotten Apple, an inside view of food integrity for professionals, policy-makers and purveyors. Subscribe for weekly insights, latest news and emerging trends in food safety, food authenticity and sustainable supply chains.

• The weird link between non-targeted food testing and protecting the Amazon rainforest

• Is training needed - the ultimate decision tool

• Blazing trails: the SOP that’s setting the supplement industry on fire

• Food Safety News and Resources Roundup (heaps of free webinars this week)

• Watermelon tastiness measurement (just for fun)

• Food fraud news, incidents and updates

I’m off to the Australian Institute of Food Science and Technology conference this morning!

Welcome to Issue 98 of The Rotten Apple, where I share how a new food test method can help save the Amazon rainforest; plus the SOP initiative that is setting the herbal ingredients world (literally) 🔥 on fire 🔥.

Shoutout to 👏Helen 👏 Russell and Stephen 👏 for upgrading to a paid subscription. Thank you for supporting my efforts to bring you interesting, ad-free food joy every week.

"Your newsletter is insightful, well researched and I like the humorous touch. What pushed me over the edge from unpaid to paid was your article on biofilms, which was a fantastic compilation of the state of knowledge. Thank you" Stephen

If you are loving this newsletter (or not), this is your last chance to provide feedback on a possible tweak to the format. Click the link below, which closes on Friday.

Have a great week,

Karen

P.S. Reminder about links and sources: some company email systems strip links from emails. If you can’t see citations or sources, which I often link to inline (like this), try reading the email directly on the website (click ‘read online’ near the top of the email).

Cover image: Vlad Hilitanu on Unsplash

---

The Weird Link Between Non-Targeted Food Testing and Protecting the Amazon Rainforest

What is non-targeted testing?

Targeted testing is the process of testing for a specific characteristic of a food, or for the presence of a specific adulterant in food.  Testing a sample of minced meat for horse DNA is targeted testing, because you are looking for the ‘target’ adulterant: horse meat.  Other examples of targeted tests include chemical analytical methods for detecting melamine in milk powder or lead-based textile dyes in spices.

Non-targeted tests do not reveal any specific characteristics of a food and do not quantify the presence of an adulterant in a test sample.  Instead, they can be used to determine whether a food sample is typical of an authentic sample or is a-typical.  Samples that are atypical could be affected by food fraud, or they could be different from ‘typical’ due to a benign reason. 

Non-targeted tests use sophisticated mathematical modelling to statistically compare the analytical ‘fingerprints’ of hundreds of known samples to the fingerprint of a test sample.  Traditional analytical methods such as mass spectrometry, stable isotope ratio analysis or gas chromatographic methods are used to create the ‘fingerprints’.

Large sets of spectra from mass spectrometric analyses are combined to create a statistical model of authentic food samples.  Left: A FT-MS spectrum (Source: Van den Brink, et al (2001))  Right: Fruit samples from different regions and growing periods can be discriminated using stable isotope ratios and statistical modelling (Source: Use of Stable Isotope Analysis in Commercial Food Authenticity Testing (2107) D. Psomiadis, (Webinar))

Because each ‘fingerprint’ contains a lot of information – for example many peaks on a spectrum - computer modelling processes are used to create a multidimensional model of dozens or hundreds of known samples.  When enough samples have been included in a model, it can be used as a comparison for unknown samples.

In food fraud testing, the set of samples used to build the model is often referred to as a ‘database’.

A major advantage of non-targeted testing for food fraud is that you do not need to know what type of food fraud or adulteration you want to test for.  This makes non-targeted testing suitable for finding fraud that was unexpected and uncovering new adulterants.

Non-targeted testing can also be used to differentiate between ‘populations’ of authentic samples – for example, agricultural commodities grown in different locations.  These population characteristics allow for an unknown test sample to be matched to a known population. 

For example, in the image above an apple grown in Slovenia in 2016 has a sufficiently different set of stable isotope ratios that it can be differentiated from an apple grown in Austria in 2015.  An apple from an unknown origin could be matched to one of these populations by testing it for stable isotopes and applying the same statistical model.  

Soybeans and the Amazon Rainforest

Soybean production has doubled since 2002 and increased tenfold since the 1960s.  Much of the increase in production has been possible because of the expansion of croplands into previously uncropped areas, including rainforests (all sources are listed at the end of this post).

Global soybean production has doubled since 2002. Source: FAO via OurWorldinData

Brazil eclipsed the USA as the world’s largest soy producer in 2019. Argentina is the other major producer with 11% of global supply.

Brazil has tripled the amount of land it uses to grow soy since 1980 and some of that land has been made available through deforestation.  More than 11 million acres of the Amazon rainforest were cleared between 2019 and 2022 and 792,051 square kilometres of rainforest have been lost since 1970.  That’s bigger than the area of France (and Texas!)

Testing for ‘rainforest’ in soybeans

Consumers and food manufacturers want to avoid soy that is contributing to deforestation but recognise that supply chains are complex and the origins of soybeans can be masked by food fraud perpetrators. 

Researchers from Queens University Belfast responded to calls from industry who requested a laboratory method that would help identify soy from problematic areas. The researchers are now in the process of developing a commercially available and ISO-accredited technique for determining soybean origin. 

They chose an elementomics technique - an untargeted test method - which makes use of the differences in concentrations of chemical elements between soybeans from different areas.  In this technique, forty-five elements are measured using Inductively Coupled Plasma Mass Spectroscopy (ICP/MS).  Two other sophisticated spectroscopic methods (LC/QToF, GC/MS/MS) were also investigated but did not provide as much differentiation between the samples as the ICP/MS method.

How to build an untargeted test method

Hundreds of samples from eight soybean-growing countries were collected for testing from contacts in each country.  Each sample was tested ten times (ten replicates!) using ICP/MS to measure how much of each of forty-five elements such as iron, selenium and sulphur, was in each sample.  

With each test run, a certified reference sample of soy (CRM) was also tested, to confirm that the instrument was performing consistently over the two years that it took to perform the tests.

The researchers then combined the results using computer-powered statistical models.

Different computer modelling processes were used to interpret the results, and the models compared to see which one was most capable of accurately identifying soybeans from different regions. The computer models were validated more than 50 times with previously-tested samples and new samples of known origin.

Source: Quinn, B (2023) (‘CRM’ = certified reference samples).

PCA (predictive discriminate analysis) and OPLS-DA (Orthogonal Projections to Latent Structures Discriminant Analysis) were two of the statistical modelling tools tested by the researchers.  PCA and OPLS-DA are techniques for analysing large datasets with high numbers of dimensions.  For the soybean elementomics project, OPLS-DA proved to be a better model, because it was able to clearly differentiate soybeans from the top five producing countries.

Thanks to the development of this method, it is now possible to perform a lab test on a soybean sample and find out which country it was grown in.  The method has an accuracy of 97.5% when determining the country of origin of soybeans from eight countries including Brazil, USA, Argentina, China and India.

Using the test to protect the rainforest

Last month, a new EU regulation on deforestation-free products came into force.  Businesses that sell commodities that are high risk for deforestation, including soy, beef, palm oil, cocoa, coffee and chocolate are now obliged to show evidence that they are not sourcing products that contributed to recent forest degradation before importing them to EU markets. 

The rules will be enforced partly by classifying the country of origin as high, standard or low risk for deforestation.  Unfortunately, the new rules could increase food fraud by motivating traders to misrepresent the geographical origin of commodities from high risk countries.

Commodity buyers and traders will be able to use the new test to make sure their soy has genuinely been grown in countries where deforestation is not a major concern, thereby reducing demand for rainforest-unfriendly soy.

What’s Next

Right now, it is not possible to distinguish between soy from recently deforested areas and older farming areas of Brazil.  However, if enough samples can be obtained from the relevant regions of Brazil, that may soon be possible too.  

The team which developed this technique is planning to make it commercially available, and it will be ISO accredited by UKAS.

In short:  🍏 Soy production contributes to deforestation in countries like Brazil 🍏 Soy supply chain stakeholders asked Queen’s University Belfast’s food fraud group to develop an analytical method for determining the origin of soybeans 🍏 An elementomics method was developed that can identify the origin of soybeans at an accuracy of 97.5% 🍏 This method will support soy purchasing companies’ efforts to avoid sourcing from areas that have been recently deforested 🍏

Sources

• Monitoring Global Soybean Production Using Elementomics to Combat Rainforest Destruction

• OurWorldinData- Soy

• The Great Organic-Food Fraud | The New Yorker

• PCA as a practical indicator of OPLS-DA model reliability - PMC (nih.gov)

• EUR-Lex - 32023R1115 - EN - EUR-Lex (Deforestation Regulation)

🍏 The main source for this post was Dr Brian Quinn’s presentation at the UK Government Chemist Conference 2023.  Watch a replay of Dr Quinn’s talk online 🍏

Can you help?  If you have access to genuine soybean samples from verified locations, Dr Brian Quinn, Head of Mass Spectrometry at QUB really (really!) wants them.  Contact details.

---

---

Blazing Trails:  How the American Botanical Council is Setting the Supplement Industry on Fire

In Issue 93 I talked to Stefan Gafner, PhD, Chief Science Officer of the American Botanical Council (ABC), about fraud and adulteration in herbs and the herbal supplements industry. In this issue, I cover the ABC’s trailblazing policy to keep adulterated botanicals out of the supply chain.

Newsflash: this initiative has just received the ‘Industry Initiative of the Year’ Award from online trade publication NutraIngredients, whose editors called it a ‘gamechanger’ and ‘monumental’.

Introducing:  Burn It Don’t Return It

The problem:  When a shipment of dodgy botanical ingredient is rejected by a reputable supplement manufacturer - perhaps because adulterants have been detected, or perhaps because of some other defect – it will be returned to the supplier, who could try to sell it on to some other, less diligent manufacturer.

The solution:  The American Botanical Council (ABC) set out to educate the industry about this issue, much as they do with their Botanical Adulterants Prevention (BAPP) bulletins.  However, a better idea soon presented itself: instead of writing an educational bulletin about the problem, they decided to create something more actionable. 

BAPP pioneered a solution to help keep fraudulent and defective materials out of the supply chain.  It includes:

• a template for a purchasing contract that helps purchasers and suppliers agree on what to do with irreparably defective materials which can be modified and customised for each organisation;

• dispute resolution and non-disclosure procedures and agreements;

• a model SOP template which can be customised to complement existing GMP programs.

The result is that any out-of-specification material that cannot safely be used for other purposes or reworked must be securely destroyed by a certified third party so that it does not end up back in commerce. This is the ‘Burn It, Don’t Return It’ initiative.

Examples and implementation: Reputable purchasers are adapting and implementing new contract clauses into supply contracts. Their suppliers then have to accommodate those requirements. 

Some suppliers are concerned that they will be forced to pay for the destruction of all out-of-spec ingredients. It is very expensive to securely destroy materials in such a way that they are rendered completely unrecoverable and to certify the destruction. 

However, the policy does not apply to all defective materials, only the few that are deemed irreparably defective.  For example, St Johns Wort extract that has been adulterated with synthetic dye and turmeric adulterated with lead chromate are both irreparably defective and also pose health hazards to consumers.  These must be securely destroyed so that they don’t return to commerce.  On the other hand, gingko leaf extract that contains too few glycosides to meet the original purchasers’ specifications could be used legitimately in a different product or formulation and so does not need to be destroyed.

Because of this, the policy only needs to be used very rarely, if at all, with legitimate suppliers who are trying to do the right thing.

I love the Burn It Don’t Return It policy: it’s a trail-blazing initiative that turns adulterated botanicals to ashes and illuminates a path towards a fraud-free future 🔥

The American Botanical Council’s Burn It Don’t Return It initiative

---

Food Safety News and Resources

Lots of good free webinars this week!

---

Chart of the Week: Watermelon (Just for Fun)

Found this delicious chart on social media the other day. 😊

Image credit: Non aesthetic things on Twitter

---

What you missed in last week’s email

• Biofilms Masterclass

• Sustainability for Laboratories

• Why Daily Harvest Used Toxic Ingredients

• Carbon Offsetting Explainer Video (NSFW!)

---

Below for paying subscribers: Food fraud news, incident reports, and emerging issues, plus 🎧 an audio version 🎧 for your listening pleasure :)

📌 Food Fraud News 📌

More than two-thirds of coconut oils (n = 16) tested during a Canadian government food fraud initiative were found to be ‘unsatisfactory’ with respect to