Like many other biological sciences, we now recognize nutrition research to be a matter of gene-environment interaction. Due to its complexity, nutrition has traditionally been an observational science; matching physiology with molecular thinking was next to impossible. As a result, animal models were discarded and human studies were limited to available biomarkers. With these blunt tools, human studies proceeded in heterogeneous patient populations.

The nature of nutrition is nuanced. Unlike pharmacologic interventions, food contains multiple bioactives, usually with low receptor affinities. Consequently, weak or even conflicting biological effects were observed. In seeking the statistical power of large cohorts, subtle differences become blurred. Several times, epidemiology and meta-analysis have killed solid evidence in properly performed human intervention trials and mechanistic research in animal models.¹

Apparently, the major dilemma that nutrition research faces is a focus on health instead of disease without the tools to quantify health properly. So, the challenge in meeting this dilemma will be to capture, quantify, and select the relevant molecular changes among the noise of healthy variation and to do this on an individual basis. Thus, nutrigenomics will require a systems biology toolkit adapted to nutritional specifics. As it stands, the commercial activities that have sprung up around this endeavor have too little to spend on R&D to really make this effort.

So, in Europe, some 20 universities have met this challenge in creating the European Nutrigenomics Organization (NuGO). With the help of ?17 million, we have embarked in a joint effort aimed at reshaping nutrition research.

The four prime objectives are to:
• Quantify personal health in relation to nutrition
• Develop a comparative nutritional systems biology
• Quantify the genetic component in the nutrition-health relationship
• Develop and implement the related technology, training, and infrastructure

In my view, this is both a timely and unique collaborative enterprise. NuGO aims not to fund actual research projects, but to build the infrastructure that creates and stimulates joined research. Examples include a nutritional metabolomics initiative, a nutrient-sensitive gene database, a shared transcriptomics pipeline including a nearly ?480,000 array deal with Affymetrix, a controlled data-sharing environment where all 800 member researchers can have access to each other's data and results without losing control and ownership, and an extensive training program for junior scientists. Some 10 new research initiatives have now started, each involving at least four NuGO partner institutes. NuGO is now in its third year and opening up to others to join this effort.

A number of pioneering commercial activities have begun to apply nutrigenetic principles, mostly through genotyping a limited number of polymorphisms in relation to dietary advice, or by selling dietary supplements. While it's unclear whether these early adopters will do more harm than good for the field, they've certainly created an awareness of the possibilities. A major problem with advice so far is that it does not effectively combine genotype and phenotype.

In my vision, relevant commercial enterprises will come only when we are able to quantify health both through genotype and phenotype. This needs to be done first in the scientific arena through controlled human nutritional intervention studies containing both components. This will create a new knowledge base for both product development and segmentation. Ultimately, healthy nutrition will be an individualized market following the trends of personalized medicine. It may just take more work and public support to get us there.

Ben van Ommen heads the nutrigenomics activities at TNO Quality of Life and is director of NuGO (www.nugo.org).