Final Report Summary - MYCOSPEC (Novel infrared spectroscopic tools for mycotoxin determination in foodstuffs for increased food safety)
Executive Summary:
The MYCOSPEC project was a two year applied research project funded under the “Research for SMEs” programme of the European Commission’s Seventh Framework Programme (FP7). The project started in November 2013 with the aim to advance the state of the art on spectroscopic technologies for the detection of mycotoxins contamination in foodstuffs and agricultural commodities to ensure food and feed quality.
Mycotoxins are contaminants produced by molds that invade and grow on virtually all food crops and processed foodstuff if conditions are favorable to fungal growth, being the cause of both acute and long term illness. This toxins greatly resist decomposition or being broken down in digestion, so when farm animals are exposed either to high levels of mycotoxins or to low levels over a longer period of time, there is a risk that significant amounts of the mycotoxins will be carried over into animal products such as milk, eggs and meat.
The impact of mycotoxin exposure depends on the toxicity of the mycotoxin, the extent of exposure, age and nutritional status of the individual and possible synergistic effects of other chemicals to which the individual is exposed.
Mycotoxin contamination is considered an unavoidable risk because the mold growth is dependent on environmental conditions. Mycotoxins have significant economic impacts in numerous crops, especially wheat, maize, peanuts and other nut crops, cottonseed, and coffee. The Food and Agriculture Organization has estimated that 25% of the world’s crops are affected by mycotoxins with important economic consequences due to 1) yield loss due to diseases induced by toxigenic fungi; 2) reduced crop value resulting from mycotoxin contamination; 3) losses in animal productivity from mycotoxin-related health problems; and 4) human health costs. Additional costs associated with mycotoxins include the cost of management at all levels - prevention, sampling, mitigation, litigation, and research costs. These economic impacts are felt all along the food and feed supply chains: crop producers, animal producers, grain handlers and distributors, processors, consumers, and society as a whole (due to health care impacts and productivity losses).
To protect consumers and to keep food and feed safe, regulatory authorities have set strict regulatory controls in most importing countries, thus affecting international trade. As a result of that, foodstuff buyers increasingly demand more rigorous and timely food safety testing tools.
However, current methodologies for mycotoxin detection usually require laborious sample preparation and laboratory testing, failing to provide rapid measurements and being costly, time-consuming and unsuitable for field application.
In recent years, the inherent limitations of current reference methods have driven the development of new analytical methods based on spectroscopic technologies. These methods are rapid, require very little labor, training and limited sample preparation. These methods are commonly utilized in food process control and have been investigated for mycotoxin contamination in a variety of cereals. However, to use this technology widely and efficiently, further improvements to mycotoxin detection limits are needed, as well as reduced variability among samples, decreased errors in the reference methods, and the development of robust calibration models.
The overall goal of the project was to develop an innovative tool based on infrared spectroscopic fingerprinting techniques for the rapid on-site diagnosis of fungal diseases in foods or liquid foodstuffs such as e.g. homogenized foodstuff or beverages. The MYCOSPEC technology will ensure that large volumes of the cereals and foodstuffs can be probed and sensitively detected for mycotoxin contamination. The high information content of infrared spectroscopic analysis combined with multivariate calibration and data classification provides a novel approach for rapid, high-throughput monitoring of plant main components obtaining rapid information on crop quality and safety essential for screening systems in the food and feed sector.
Project Context and Objectives:
The overall goal of the MYCOSPEC project is to develop a pre-competitive system based on infrared spectroscopic fingerprinting technology for enabling efficient on-site diagnosis of mycotoxin contamination in feed, foods or in liquid foodstuffs.
In order to achieve the above, the specific objectives were as follows:
1. To commence with a clear definition of the system specifications aligned with SME requirements, as well as industry regulations in terms of food screening technology and processes for mycotoxin determination, and to use the findings to define the specifications for the MYCOSPEC system.
2. To carry out laboratory trials with QCLs and thin-film waveguide technology in order to identify QCL accessible spectral windows for indirect and direct detection of target contaminants.
3. To carry out laboratory trials with IR spectroscopy in order to build a database of clean and contaminated samples for statistical analysis. Implementation of a multivariate data analysis software package.
4. To design and build a precompetitive MYCOSPEC system.
5. To integrate the system hardware with the software developed in WP3 and to develop an easy to use process control that will enable the MYCOSPEC system, and will ensure that the MYCOSPEC system is flexible enough to be readily operable with existing process control software (WP4).
6. To install and set up the system at the industry sites, to ensure the system is fully operational and that the SMEs are capacitated to use prototype.
7. To carry out trials to validate and demonstrate the performance of the prototype system in the field for determining mycotoxins in cereals, peanuts, apple juice and beer and to carry out laboratory analysis to cross-reference with conventional analysis methods in order to validate the ability of MYCOSPEC prototype to accurately screen grains, foodstuffs and beverages in industry.
8. To carefully outline scaling-up rules and development work towards a fully commercial system so that the SMEs are provided with the knowledge that they need to exploit the results in the market.
9. To facilitate the uptake of the MYCOSPEC results by the participating SMEs of the consortium, as well as a wider audience of SMEs and stakeholders from the food supply and value chains, by carrying out a comprehensive series of knowledge transfer and training activities to ensure that knowledge of generated results is transferred to the SMEs, and that they (and their clients) are trained on the use and benefits of the technology.
Project Results:
With the aim to develop an innovative tool based on infrared spectroscopic fingerprinting techniques for the rapid on-site detection of mycotoxin contaminants in foodstuff. The first target of the project focused on understanding the needs of industrial stakeholders (European cereal, beverage and food producers and suppliers) to ensure that the MYCOSCPEC system developed over the course of the project could be exploited by industry and would have a far-reaching impact in the marketplace. To this end, both the industry and research partners worked closely in developing questionnaires that served as a basis for direct consultations, which were conducted via either on-site visits or phone interviews with industries of the sector, in order to define the industrial specifications that the MYCOSPEC system should meet, as well as market needs and perceptions in terms of product quality, new technologies, etc.
The overall MYCOSPEC architecture was defined and was used as a blueprint for guiding the technical work carried out during the project. Briefly, MYCOSPEC would be based on mid-infrared spectroscopy using tunable QCL technology and thin-film waveguides for rapid detection of fungal infection in food. It was agreed that the industrial prototype would be a static, bench top instrument to be used off-line at any stage of the production process, enclosed in a robust packaged suitable for use in a dusty environment but with easy access to the waveguide and able to detect fungal infection in different matrixes.
As part of technical work UULM established a reproducible micro-fabrication protocol for GaAs/AlGaAs waveguides, suitable for various waveguide shapes and dimensions. Additionally, a protocol for quality control of the obtained waveguide chips was implemented to ensure reproducibility and continuous production of high-quality waveguide devices. The waveguide structure was optimized to maximize the obtained absorbance signals. To this end, prospective spectral simulations of the GaAs/AlGaAs waveguide structure were conducted. Experimentally, the MIRcat system was used, which combines several QCLs into one compact device to enable access to the entire MIR fingerprint region of interest. Based on broadband IR studies, two spectral regions of interest (the amide spectral range and the carbohydrate region were determined as providing most spectral information for fungal contamination detection, and have thus enabled selecting two tunable QCLs covering the respective spectral regions for establishing the field-deployable MYCOSPEC prototype. Based on the selected spectral regions, simulations were performed for determining the optimum waveguide width for straight strip waveguide structures. Finally, it was demonstrated for solid as well as liquid (extract) samples that the combination QCL/thin-film waveguide provides superior results compared to conventional IR-ATR using broadband FTIR methods. Finally, slab waveguides with a 6 µm thick GaAs core waveguiding layer were determined as the optimum choice to fulfil industrial application requirements including cost efficiency, potential for mass fabrication, ease of handling, and optical coupling efficiency. Furthermore, a modular waveguide mount was designed, fabricated, and tested suitable for solid as well as liquid phase measurements.
In parallel to the previous work, BOKU started the work by identifying MIR spectral window for the FTIR- attenuated total reflection (ATR) measurements of commodities identified by the small and medium enterprises (SMEs) from the Consortium. Large sets of spectral data were generated from a variety of solid (maize, wheat and peanuts) and liquid (apple juice) commodities. Additionally mycotoxin concentration was determined by validated LC-MS/MS methods in order to correlate the MIR spectral information with a reliable mycotoxin concentration and to develop appropriate chemometric models that would enable the classification of contaminated and non-contaminated samples. A spectral database was created for each commodity to be used as a training (and test) set for the model to ensure accuracy and model stability. In summary, samples from 2 solid food commodities (maize and peanuts) were successfully classified at their respective legal limits (500 µg kg-1 and 8 µg kg-1) using a decision tree-based classification method (“Bagged Trees Classification”). Cross validation was performed on the data and validation accuracies between 77 and 95% were achieved. Apple juice samples were successfully classified at a concentration of 1000 µg kg-1. It could be concluded that an IR spectral window between 800 and 1800 cm-1 used for chemometric modelling provides suitable information for classification based on spectral changes induced by fungal infection. Results provided a sound database for future QCL measurements. Spectral windows covered by the laser-based system were included in the evaluation of the spectra.
Based on the system specifications defined, as well as knowledge acquired in previous work, hardware and software was designed along WP4 for developing MYCOSPEC prototype which intend to be a novel system based on tunable QCL-based infrared spectroscopic and thin-film waveguides technologies for the rapid on-site detection of mycotoxin contaminants in foodstuff. Thanks to QCL-based systems it is possible to reduce the size to an on-chip level while providing superior emission power, reliability, wavelength tunability, and lifetime. The design involves the integration of the components developed and chosen during previous work: waveguide and its holder, as well as the detector unit, MIRcat system, data pre-processing and chemometric models, by keeping the specifications. The final prototype is comprised of three main function blocks: optics, electronics and software enclosed in a robust package suitable for use in a dusty environment with two chambers. The sample chamber integrates a sample press and can be accessed by any user. It is isolated from the main enclosure, which contains all the optics and electronics and is only accessible by an Admin user. In addition, to the optical, opto-mechanical and mechanical design of the prototype, a suitable electronic detection scheme was developed and environment control and safety features such as dehumidification and interlocks were also implemented. Furthermore, a control board was developed and built to implement the detection scheme and allow communication between hardware and software. The user interface, software and process control, was then developed and integrated into the system hardware. Based on the design parameters, all the elements were built and integrated into a single unit that is fully controlled through the Mycospec software. The implementation of multivariate data analysis methodology into a software package provide an immediate YES/NO indicator for final decision making. Finally the prototype underwent functionality tests to ensure it performed as expected and to find a suitable overall setup for carrying out the validation trials.
The validation of the system started with a preliminary validation of the final system at UULM with the support of all partners. The feasibility of transferring QCL spectra into a FTIR spectral database was successfully demonstrated by developing Chemometric models individually for both systems FTIR and QCL, and also for combined PCA models (i.e. transferability of calibration/classification model). The chemometric models obtained were adjusted to the prototype setup in order to enable a more straightforward validation during the last months of the project. The results showed that the MYCOSPEC system can effectively be a tool for detecting on-site mycotoxin contamination above legal limits in foodstuffs, by achieving excellent separation of validated contaminated and uncontaminated maize, peanut and wheat samples by using the MYCOSPEC system. An overall improvement of mycotoxin analysis was shown with the MYCOSPEC system compared to the FTIR measurements reported in previous workpackages, in particular the limit of separation for peanuts was improved from 8 μg/kg to 2 μg/kg for Aflatoxin B1 and it was possible a separation of DON contaminated wheat samples at EU regulatory limits by first time.
Potential Impact:
The food supply chain is currently a major contributor to the European economy, connecting the agricultural, food processing and distribution sectors that together constitute more than 7% of European employment. About 25% of the world´s food supply is lost due to microbial spoilage during production processes and storage. Thus, spoilers have a huge economic and environmental impact. It has been reported that in the EU around 40 million tons of food are wasted each year during food manufacturing and whole sale storage, with estimated loss of value of around 35 billion €. In addition to food waste, microbial food spoilers create enormous health problems, with estimated costs for the health system in the EU of around 7 billion € per year.
In particular, mycotoxins enter in the food chain by infecting crops that are directly consumed by humans or indirectly as feed ingredients ingested by animals which can be metabolized and remain in milk, meat and eggs. The fungus that produce such mycotoxins can be reduced by proper drying, sorting and storage however it is not possible to impose a total ban of such contaminants because they occur naturally. Due to its high toxicity, the most of countries have introduced legislation concerning mycotoxins and have imposed maximum admissible levels of in commodities.
With growing concerns for food safety, the market for food safety testing products has been witnessing steady growth over the past several years. Major market drivers are globalisation of food supply trade and growing concerns for food safety and increasing regulations. Other factors propelling market growth include increased demand for food testing equipment with high-end technology and speedy performance, and a shift towards more clear and regular communication of food safety standards.
The MYCOSPEC project has developed a pre-competitive system based infrared spectroscopic fingerprinting techniques and novel IR-laser and waveguide technologies for the rapid on-site diagnosis of fungal contaminations in foods and in liquid foodstuffs. The high information content of infrared spectroscopic analysis combined with multivariate calibration and data classification provides a rapid, high-throughput monitoring tool for inspecting crop quality and safety in the food and feed sector. Data acquired is stored in detail (i.e. on a single spectra basis) for validation exercises with established lab-based methodology, which acts as a reference point for field-based measurements to document accuracy and precision of the obtained analytical results. Implementation of multivariate data analysis methodology in an easy-to-use software package provide an immediate yes/no indicator to non-expert personnel for final decision-making. Both, measurement and data analysis approaches are packaged in a rugged system suitable for deployment in agro-industrial environments.
The intention of the proposed novel IR-spectroscopic approach is not to replace established usually laboratory-bound-high performance analytical methods for mycotoxin determination, such as LC-MS, but to provide a useful complementary technical platform to quickly identify mouldy cereals, nuts and derived foodstuffs and to estimate the concentration of mycotoxins in the investigated food commodities applicable in a variety of field environments.
The knowledge generated throughout the project was effectively transferred to SMEs of the Consortium through comprehensive knowledge transfer activities to ensure a profitable knowledge transfer from the RTDs to the technical and managerial staff of the industrial partners of the Consortium. Training activities consisted of a complementary approach based on written documentation, power point presentation, one on-site training sessions for the Consortium members. Such activities were fundamental to facilitate acceptance and future uptake and exploitation of the results of the project by the participant SMEs.
In terms of Dissemination, the public website of the project (http://www.mycospec.eu) informs to the public and interested parties about latest news and the progress of the project. Various press releases were circulated in professional food technology magazines and websites, raising the awareness of MYCOSPEC technology both in industry and in the public domain. Moreover, the technology was demonstrated internally to the consortium members at different stages of development. The project partners also took part in number of trade shows and exhibitions such as Alimentaria (an International Trade Fair in food and beverages which closed the 2014 edition with 140.000 visitors, 30% of them from other countries and 3.800 exhibitors) or BTA 2015-Barcelona Food Technologies which offers the latest innovative and technology solutions for the food industry. In addition, the main technical results were presented in technical Conferences such as the 8th conference of The World Mycotoxin Forum, which is the leading international meeting series on mycotoxins. Finally, 1 scientific article was produced and published at the scientific journal Analitycal and Bionalitical Chemistry under the title “Advancements in IR spectroscopic approaches for the determination of fungal derivades contamination in food crops”, and it is envisaged to publish at least two new scientific articles in peer reviewed publications during post-project phase. Overall, 53 dissemination activities have been carried out throughout 27 months of the project.
In terms of exploitation, the SMEs understanding is that further development is required to reach a saleable state. In fact, at the end of the project a novel tool for mycotoxin detection based on QCl and waguideline technology for rapid mycotoxin detection has been developed and validated for wheat, peanuts and maize, however further developments is required to expand the number of matrices that can be analyzed since the process of preparing fully characterized samples and acquiring the corresponding spectra is time consuming. At this point, it is clear for SMEs that further development is required for both, the development of new models and improvement of equipment performance to reach a saleable state. The SMEs could look together for new funding opportunities within the H2020 2016-2017 work programme but in the meantime they have agreed that prototype remain at ULM for continuous validation.
Once the system reach a commercializable state, is intend that the participant MGOS commercialize the MYCOSPEC system in the marketplace which involve manufacturing, marketing, sales, installation and maintenance of the system. Furthermore (ICC, FM, CLG) and the Union of producers SETBIR, as end users will have preferential use and they could uptake the MYCOSPEC system in their own facilities to ensure safety of their foods suppliers and essentially acting as early adopters or first clients of the system, thus benefiting from the preferential use conditions
List of Websites:
www.mycospec.eu
Dr Edurne Gaston Estanga
IRIS SL
+34935570111
egaston@ris.cat
The MYCOSPEC project was a two year applied research project funded under the “Research for SMEs” programme of the European Commission’s Seventh Framework Programme (FP7). The project started in November 2013 with the aim to advance the state of the art on spectroscopic technologies for the detection of mycotoxins contamination in foodstuffs and agricultural commodities to ensure food and feed quality.
Mycotoxins are contaminants produced by molds that invade and grow on virtually all food crops and processed foodstuff if conditions are favorable to fungal growth, being the cause of both acute and long term illness. This toxins greatly resist decomposition or being broken down in digestion, so when farm animals are exposed either to high levels of mycotoxins or to low levels over a longer period of time, there is a risk that significant amounts of the mycotoxins will be carried over into animal products such as milk, eggs and meat.
The impact of mycotoxin exposure depends on the toxicity of the mycotoxin, the extent of exposure, age and nutritional status of the individual and possible synergistic effects of other chemicals to which the individual is exposed.
Mycotoxin contamination is considered an unavoidable risk because the mold growth is dependent on environmental conditions. Mycotoxins have significant economic impacts in numerous crops, especially wheat, maize, peanuts and other nut crops, cottonseed, and coffee. The Food and Agriculture Organization has estimated that 25% of the world’s crops are affected by mycotoxins with important economic consequences due to 1) yield loss due to diseases induced by toxigenic fungi; 2) reduced crop value resulting from mycotoxin contamination; 3) losses in animal productivity from mycotoxin-related health problems; and 4) human health costs. Additional costs associated with mycotoxins include the cost of management at all levels - prevention, sampling, mitigation, litigation, and research costs. These economic impacts are felt all along the food and feed supply chains: crop producers, animal producers, grain handlers and distributors, processors, consumers, and society as a whole (due to health care impacts and productivity losses).
To protect consumers and to keep food and feed safe, regulatory authorities have set strict regulatory controls in most importing countries, thus affecting international trade. As a result of that, foodstuff buyers increasingly demand more rigorous and timely food safety testing tools.
However, current methodologies for mycotoxin detection usually require laborious sample preparation and laboratory testing, failing to provide rapid measurements and being costly, time-consuming and unsuitable for field application.
In recent years, the inherent limitations of current reference methods have driven the development of new analytical methods based on spectroscopic technologies. These methods are rapid, require very little labor, training and limited sample preparation. These methods are commonly utilized in food process control and have been investigated for mycotoxin contamination in a variety of cereals. However, to use this technology widely and efficiently, further improvements to mycotoxin detection limits are needed, as well as reduced variability among samples, decreased errors in the reference methods, and the development of robust calibration models.
The overall goal of the project was to develop an innovative tool based on infrared spectroscopic fingerprinting techniques for the rapid on-site diagnosis of fungal diseases in foods or liquid foodstuffs such as e.g. homogenized foodstuff or beverages. The MYCOSPEC technology will ensure that large volumes of the cereals and foodstuffs can be probed and sensitively detected for mycotoxin contamination. The high information content of infrared spectroscopic analysis combined with multivariate calibration and data classification provides a novel approach for rapid, high-throughput monitoring of plant main components obtaining rapid information on crop quality and safety essential for screening systems in the food and feed sector.
Project Context and Objectives:
The overall goal of the MYCOSPEC project is to develop a pre-competitive system based on infrared spectroscopic fingerprinting technology for enabling efficient on-site diagnosis of mycotoxin contamination in feed, foods or in liquid foodstuffs.
In order to achieve the above, the specific objectives were as follows:
1. To commence with a clear definition of the system specifications aligned with SME requirements, as well as industry regulations in terms of food screening technology and processes for mycotoxin determination, and to use the findings to define the specifications for the MYCOSPEC system.
2. To carry out laboratory trials with QCLs and thin-film waveguide technology in order to identify QCL accessible spectral windows for indirect and direct detection of target contaminants.
3. To carry out laboratory trials with IR spectroscopy in order to build a database of clean and contaminated samples for statistical analysis. Implementation of a multivariate data analysis software package.
4. To design and build a precompetitive MYCOSPEC system.
5. To integrate the system hardware with the software developed in WP3 and to develop an easy to use process control that will enable the MYCOSPEC system, and will ensure that the MYCOSPEC system is flexible enough to be readily operable with existing process control software (WP4).
6. To install and set up the system at the industry sites, to ensure the system is fully operational and that the SMEs are capacitated to use prototype.
7. To carry out trials to validate and demonstrate the performance of the prototype system in the field for determining mycotoxins in cereals, peanuts, apple juice and beer and to carry out laboratory analysis to cross-reference with conventional analysis methods in order to validate the ability of MYCOSPEC prototype to accurately screen grains, foodstuffs and beverages in industry.
8. To carefully outline scaling-up rules and development work towards a fully commercial system so that the SMEs are provided with the knowledge that they need to exploit the results in the market.
9. To facilitate the uptake of the MYCOSPEC results by the participating SMEs of the consortium, as well as a wider audience of SMEs and stakeholders from the food supply and value chains, by carrying out a comprehensive series of knowledge transfer and training activities to ensure that knowledge of generated results is transferred to the SMEs, and that they (and their clients) are trained on the use and benefits of the technology.
Project Results:
With the aim to develop an innovative tool based on infrared spectroscopic fingerprinting techniques for the rapid on-site detection of mycotoxin contaminants in foodstuff. The first target of the project focused on understanding the needs of industrial stakeholders (European cereal, beverage and food producers and suppliers) to ensure that the MYCOSCPEC system developed over the course of the project could be exploited by industry and would have a far-reaching impact in the marketplace. To this end, both the industry and research partners worked closely in developing questionnaires that served as a basis for direct consultations, which were conducted via either on-site visits or phone interviews with industries of the sector, in order to define the industrial specifications that the MYCOSPEC system should meet, as well as market needs and perceptions in terms of product quality, new technologies, etc.
The overall MYCOSPEC architecture was defined and was used as a blueprint for guiding the technical work carried out during the project. Briefly, MYCOSPEC would be based on mid-infrared spectroscopy using tunable QCL technology and thin-film waveguides for rapid detection of fungal infection in food. It was agreed that the industrial prototype would be a static, bench top instrument to be used off-line at any stage of the production process, enclosed in a robust packaged suitable for use in a dusty environment but with easy access to the waveguide and able to detect fungal infection in different matrixes.
As part of technical work UULM established a reproducible micro-fabrication protocol for GaAs/AlGaAs waveguides, suitable for various waveguide shapes and dimensions. Additionally, a protocol for quality control of the obtained waveguide chips was implemented to ensure reproducibility and continuous production of high-quality waveguide devices. The waveguide structure was optimized to maximize the obtained absorbance signals. To this end, prospective spectral simulations of the GaAs/AlGaAs waveguide structure were conducted. Experimentally, the MIRcat system was used, which combines several QCLs into one compact device to enable access to the entire MIR fingerprint region of interest. Based on broadband IR studies, two spectral regions of interest (the amide spectral range and the carbohydrate region were determined as providing most spectral information for fungal contamination detection, and have thus enabled selecting two tunable QCLs covering the respective spectral regions for establishing the field-deployable MYCOSPEC prototype. Based on the selected spectral regions, simulations were performed for determining the optimum waveguide width for straight strip waveguide structures. Finally, it was demonstrated for solid as well as liquid (extract) samples that the combination QCL/thin-film waveguide provides superior results compared to conventional IR-ATR using broadband FTIR methods. Finally, slab waveguides with a 6 µm thick GaAs core waveguiding layer were determined as the optimum choice to fulfil industrial application requirements including cost efficiency, potential for mass fabrication, ease of handling, and optical coupling efficiency. Furthermore, a modular waveguide mount was designed, fabricated, and tested suitable for solid as well as liquid phase measurements.
In parallel to the previous work, BOKU started the work by identifying MIR spectral window for the FTIR- attenuated total reflection (ATR) measurements of commodities identified by the small and medium enterprises (SMEs) from the Consortium. Large sets of spectral data were generated from a variety of solid (maize, wheat and peanuts) and liquid (apple juice) commodities. Additionally mycotoxin concentration was determined by validated LC-MS/MS methods in order to correlate the MIR spectral information with a reliable mycotoxin concentration and to develop appropriate chemometric models that would enable the classification of contaminated and non-contaminated samples. A spectral database was created for each commodity to be used as a training (and test) set for the model to ensure accuracy and model stability. In summary, samples from 2 solid food commodities (maize and peanuts) were successfully classified at their respective legal limits (500 µg kg-1 and 8 µg kg-1) using a decision tree-based classification method (“Bagged Trees Classification”). Cross validation was performed on the data and validation accuracies between 77 and 95% were achieved. Apple juice samples were successfully classified at a concentration of 1000 µg kg-1. It could be concluded that an IR spectral window between 800 and 1800 cm-1 used for chemometric modelling provides suitable information for classification based on spectral changes induced by fungal infection. Results provided a sound database for future QCL measurements. Spectral windows covered by the laser-based system were included in the evaluation of the spectra.
Based on the system specifications defined, as well as knowledge acquired in previous work, hardware and software was designed along WP4 for developing MYCOSPEC prototype which intend to be a novel system based on tunable QCL-based infrared spectroscopic and thin-film waveguides technologies for the rapid on-site detection of mycotoxin contaminants in foodstuff. Thanks to QCL-based systems it is possible to reduce the size to an on-chip level while providing superior emission power, reliability, wavelength tunability, and lifetime. The design involves the integration of the components developed and chosen during previous work: waveguide and its holder, as well as the detector unit, MIRcat system, data pre-processing and chemometric models, by keeping the specifications. The final prototype is comprised of three main function blocks: optics, electronics and software enclosed in a robust package suitable for use in a dusty environment with two chambers. The sample chamber integrates a sample press and can be accessed by any user. It is isolated from the main enclosure, which contains all the optics and electronics and is only accessible by an Admin user. In addition, to the optical, opto-mechanical and mechanical design of the prototype, a suitable electronic detection scheme was developed and environment control and safety features such as dehumidification and interlocks were also implemented. Furthermore, a control board was developed and built to implement the detection scheme and allow communication between hardware and software. The user interface, software and process control, was then developed and integrated into the system hardware. Based on the design parameters, all the elements were built and integrated into a single unit that is fully controlled through the Mycospec software. The implementation of multivariate data analysis methodology into a software package provide an immediate YES/NO indicator for final decision making. Finally the prototype underwent functionality tests to ensure it performed as expected and to find a suitable overall setup for carrying out the validation trials.
The validation of the system started with a preliminary validation of the final system at UULM with the support of all partners. The feasibility of transferring QCL spectra into a FTIR spectral database was successfully demonstrated by developing Chemometric models individually for both systems FTIR and QCL, and also for combined PCA models (i.e. transferability of calibration/classification model). The chemometric models obtained were adjusted to the prototype setup in order to enable a more straightforward validation during the last months of the project. The results showed that the MYCOSPEC system can effectively be a tool for detecting on-site mycotoxin contamination above legal limits in foodstuffs, by achieving excellent separation of validated contaminated and uncontaminated maize, peanut and wheat samples by using the MYCOSPEC system. An overall improvement of mycotoxin analysis was shown with the MYCOSPEC system compared to the FTIR measurements reported in previous workpackages, in particular the limit of separation for peanuts was improved from 8 μg/kg to 2 μg/kg for Aflatoxin B1 and it was possible a separation of DON contaminated wheat samples at EU regulatory limits by first time.
Potential Impact:
The food supply chain is currently a major contributor to the European economy, connecting the agricultural, food processing and distribution sectors that together constitute more than 7% of European employment. About 25% of the world´s food supply is lost due to microbial spoilage during production processes and storage. Thus, spoilers have a huge economic and environmental impact. It has been reported that in the EU around 40 million tons of food are wasted each year during food manufacturing and whole sale storage, with estimated loss of value of around 35 billion €. In addition to food waste, microbial food spoilers create enormous health problems, with estimated costs for the health system in the EU of around 7 billion € per year.
In particular, mycotoxins enter in the food chain by infecting crops that are directly consumed by humans or indirectly as feed ingredients ingested by animals which can be metabolized and remain in milk, meat and eggs. The fungus that produce such mycotoxins can be reduced by proper drying, sorting and storage however it is not possible to impose a total ban of such contaminants because they occur naturally. Due to its high toxicity, the most of countries have introduced legislation concerning mycotoxins and have imposed maximum admissible levels of in commodities.
With growing concerns for food safety, the market for food safety testing products has been witnessing steady growth over the past several years. Major market drivers are globalisation of food supply trade and growing concerns for food safety and increasing regulations. Other factors propelling market growth include increased demand for food testing equipment with high-end technology and speedy performance, and a shift towards more clear and regular communication of food safety standards.
The MYCOSPEC project has developed a pre-competitive system based infrared spectroscopic fingerprinting techniques and novel IR-laser and waveguide technologies for the rapid on-site diagnosis of fungal contaminations in foods and in liquid foodstuffs. The high information content of infrared spectroscopic analysis combined with multivariate calibration and data classification provides a rapid, high-throughput monitoring tool for inspecting crop quality and safety in the food and feed sector. Data acquired is stored in detail (i.e. on a single spectra basis) for validation exercises with established lab-based methodology, which acts as a reference point for field-based measurements to document accuracy and precision of the obtained analytical results. Implementation of multivariate data analysis methodology in an easy-to-use software package provide an immediate yes/no indicator to non-expert personnel for final decision-making. Both, measurement and data analysis approaches are packaged in a rugged system suitable for deployment in agro-industrial environments.
The intention of the proposed novel IR-spectroscopic approach is not to replace established usually laboratory-bound-high performance analytical methods for mycotoxin determination, such as LC-MS, but to provide a useful complementary technical platform to quickly identify mouldy cereals, nuts and derived foodstuffs and to estimate the concentration of mycotoxins in the investigated food commodities applicable in a variety of field environments.
The knowledge generated throughout the project was effectively transferred to SMEs of the Consortium through comprehensive knowledge transfer activities to ensure a profitable knowledge transfer from the RTDs to the technical and managerial staff of the industrial partners of the Consortium. Training activities consisted of a complementary approach based on written documentation, power point presentation, one on-site training sessions for the Consortium members. Such activities were fundamental to facilitate acceptance and future uptake and exploitation of the results of the project by the participant SMEs.
In terms of Dissemination, the public website of the project (http://www.mycospec.eu) informs to the public and interested parties about latest news and the progress of the project. Various press releases were circulated in professional food technology magazines and websites, raising the awareness of MYCOSPEC technology both in industry and in the public domain. Moreover, the technology was demonstrated internally to the consortium members at different stages of development. The project partners also took part in number of trade shows and exhibitions such as Alimentaria (an International Trade Fair in food and beverages which closed the 2014 edition with 140.000 visitors, 30% of them from other countries and 3.800 exhibitors) or BTA 2015-Barcelona Food Technologies which offers the latest innovative and technology solutions for the food industry. In addition, the main technical results were presented in technical Conferences such as the 8th conference of The World Mycotoxin Forum, which is the leading international meeting series on mycotoxins. Finally, 1 scientific article was produced and published at the scientific journal Analitycal and Bionalitical Chemistry under the title “Advancements in IR spectroscopic approaches for the determination of fungal derivades contamination in food crops”, and it is envisaged to publish at least two new scientific articles in peer reviewed publications during post-project phase. Overall, 53 dissemination activities have been carried out throughout 27 months of the project.
In terms of exploitation, the SMEs understanding is that further development is required to reach a saleable state. In fact, at the end of the project a novel tool for mycotoxin detection based on QCl and waguideline technology for rapid mycotoxin detection has been developed and validated for wheat, peanuts and maize, however further developments is required to expand the number of matrices that can be analyzed since the process of preparing fully characterized samples and acquiring the corresponding spectra is time consuming. At this point, it is clear for SMEs that further development is required for both, the development of new models and improvement of equipment performance to reach a saleable state. The SMEs could look together for new funding opportunities within the H2020 2016-2017 work programme but in the meantime they have agreed that prototype remain at ULM for continuous validation.
Once the system reach a commercializable state, is intend that the participant MGOS commercialize the MYCOSPEC system in the marketplace which involve manufacturing, marketing, sales, installation and maintenance of the system. Furthermore (ICC, FM, CLG) and the Union of producers SETBIR, as end users will have preferential use and they could uptake the MYCOSPEC system in their own facilities to ensure safety of their foods suppliers and essentially acting as early adopters or first clients of the system, thus benefiting from the preferential use conditions
List of Websites:
www.mycospec.eu
Dr Edurne Gaston Estanga
IRIS SL
+34935570111
egaston@ris.cat