Using AI to speed up evidence-based product development [Interview]

“We have developed different AI modules for projects within F&B industry to identify ingredients that are beneficial for your health and/or to identify (novel) benefits of a company’s ingredient portfolio. Moreover, we apply AI to optimize (pre)clinical study strategies (biomarker discovery, assay identification, dosing optimization).”

“For example, recently we applied AI to identify biomarkers that are suitable to monitor the immune effects of immunonutrition in early life (first thousand days of life) (Van Bilsen et al., Front Immunol. 2020 Apr 17;11:644. doi: 10.3389/fimmu.2020.00644).”

“Selecting and developing the right ingredients/products starts with scientific underpinning of relevant targets/processes and read-outs (biomarkers) based on already available knowledge/data on immune health. AI can help to collect, organize and understand the structure and dynamics of our biological immune network throughout different life stages and to safely intervene in this network. Based on the outcomes, ingredients can be selected and tested with an optimized (pre)clinical study strategy.”

“Based on data-driven analysis of product-specific data (e.g. omics data), we focus on how the product affects the balances of our biological immune network, resulting in the elucidation of the mode of action of the product and a prediction on beneficial immune processes (e.g. gut barrier optimization, tolerance induction) and/or on immune disorders (e.g. colitis, allergies, dermatitis, lung inflammation etc).”

“The health claim will need to be based on evidence derived from clinical studies. TNO’s AI tools can help companies by elucidation the mode of action of their product and by optimizing their (pre)clinical study strategies (appropriate biomarkers and relevant immune health endpoints). However, although AI may help accelerate product health claim development, at present and in the near future AI tools will not be able to replace the (pre-)clinical R&D trajectories for F&B companies.”

“AI and data science require a high degree of expertise to analyze the data and derive meaningful insights. Unless AI is applied to study data derived from fit-for-purpose (pre-)clinical studies that already address the health benefits of interest, one always has to establish the desired health effects in relevant follow-on studies. AI can support in substantiating health benefits of interest and direct identification of novel health benefits, as well as support selection of relevant models and optimization of study designs, but is unlikely to replace all those in the near future.”

“At TNO, we address a wide range of aspects of relevance to human health. The immune health program links strongly to many of our in vitro models on microbiome composition and functionality, and functional preclinical tissue models such as organ-on-chip and ex-vivo intestine models. In addition, we have a strong position in assessment of metabolic health benefits. In this domain we offer both preclinical models ranging from tissue to animal models as well as in human studies, using the metabolic challenge test, i.e. PhenFlex technology. Finally, we support companies in studying health benefits in the mental health and cognition domain, by applying our in-depth expertise on body-brain interactions in combination with state-of-the-art testing strategies and technologies. Of relevance to the set-up of human studies, we have developed strong knowledge of the application of eHealth and wearable technologies to support studies in different study designs ranging from a research at a clinical site to Do-It-Yourself real world studies. All this is supported by state-of-the-art in-house data warehousing, management and analytical skills. “

“The COVID-19 pandemic has emphasized the importance of good immune health and biological resilience to external threats like the corona virus. Therefore, we expect that the coming years we will see the F&B industry focus strongly on health improvement by nutrition, especially addressing immune support. Novel technologies, including AI will strongly accelerate this field.”

“Furthermore, we expect that evidence-based personalized nutrition will become more and more and more important, particularly in areas like immune health, gut health and the microbiome. In 3-5 years truly personalized products will still be difficult to deliver as they’re not cost-effective as the field of personalized nutrition is still in its infancy. However, we expect that individuals will increasingly want to choose products fitting with their individual lifestyle goals. In order to develop such products, AI approaches for data-based personalized nutrition advice will become more and more important.”