The Periodic Table of Food Initiative (PTFI) has set itself an ambitious task: to build a global resource capable of capturing far more of food’s molecular composition than conventional nutrient databases allow. Achieving that goal, however, depends not only on analytical breadth, but on methods that can produce comparable data across foods, laboratories, and regions.
Jessica Prenni, Professor of Biochemistry and Molecular Biology, and Director of the Proteomics and Metabolomics Facility at Colorado State University, USA, is helping to establish that analytical foundation. Her lab is developing standardized, non-targeted omics workflows that can be applied consistently across the extraordinary diversity of food matrices.
Here, Prenni discusses the analytical and logistical challenges of the initiative, the compromises required to make methods globally accessible, and the long-term potential of a shared food composition resource.
What is the overall mission of the Periodic Table of Food Initiative?
The goal of the PTFI is to produce the next generation of food composition data. A major part of that is communicating and distributing standardized methods to labs around the world, so that we can work together to populate a single database that can serve as a global, gold standard resource.
The initiative is also about creating methods that enable people to generate comparable data, while building an ecosystem of researchers working toward a common goal. At this point, we have around 25 labs around the world that are either already using these methods or are in the process of implementing them.
We also have a fellows program, where we support students using these methods to answer research questions, alongside educational and training materials. Creating the data is essential, but the resource can only reach its full potential if there is also a community that is aware of it, using it, and contributing to it.
It is important to emphasize that this is not meant to replace the progress made over the past few decades. Targeted approaches focused on important macro- and micronutrients are incredibly valuable, and what we are doing is complementary. We already know a lot about those compounds, but there is a whole other world of compounds in food that we do not yet understand.
The question is how we focus on that space and add this new knowledge to what we already know. It is very much about expansion, not replacement.
Can you tell us about the focus of your current research in this area?
In the simplest terms, we are part of a larger team developing standardized omics methods that can be applied across many different food matrices. Over the past few years, my lab has become increasingly involved in the project, to the point where it is now our primary focus.
Food is the most difficult matrix I have ever worked with. Every food is different, often drastically so: a fatty, protein-rich meat product presents very different challenges from a water-rich fruit or vegetable. That creates a host of analytical challenges, and one of the biggest is identifying methods that work well across matrices.
There are many labs around the world performing omics analysis of food, but this matrix effect is a major challenge. A lab may decide to focus on one specific food type, tailoring its extraction and analytical workflows to that specific material. That can generate excellent data, but if another lab analyzes the same food under even slightly different conditions, it becomes very difficult to compare results across studies.
Our goal is to create a data resource that can eventually be leveraged using emerging AI tools to mine data, make predictions, and build stronger links between areas such as climate change and food composition, or food composition and human health. To do that, we need comparable data at scale. We already have the omics technologies, but if everyone has their own individual approach, meaningful comparison across studies becomes challenging.
That is where my lab’s focus has been as part of this larger team: developing standardized methods. In a sense, we are actually dialing some methods back. They may not be the absolute optimal methods for analyzing one specific food, but they work well across many foods. The idea is to create these methods, validate them, and then apply them broadly.
From your perspective, what makes this initiative significant for the field of food analysis, and what kinds of analytical or data challenges has it revealed so far?
Nothing we are doing analytically is especially groundbreaking or novel in itself. We are simply leveraging what the field has already developed in omics and working to make those methods more generalizable.
The biggest challenge is determining how to design a method that works across all of these different matrices. That is difficult, but not impossible. The bigger issue is that this initiative involves working with labs all over the world, including in high-, middle-, and low-income countries. When developing these methods, we have to carefully consider the technologies, reagents, and consumables we are using – and whether they are actually accessible to labs everywhere.
That has been one of our biggest challenges. We have partner labs in low- and middle-income countries in regions such as Africa and South America, where something as simple as accessing solvents to run assays can cost significantly more and take months to arrive.
The logistics of getting what you need, when you need it, can be far more difficult than we ever anticipated. As a result, we have to be really mindful about our choices. One solvent may seem more appropriate analytically, but if it is not easily accessible around the world, then it is not the right choice. In many cases, we need to use something more widely available, such as methanol, because it is easier for researchers to obtain.
All that said, I have found it incredibly rewarding to work with partners around the world – and to understand what foods they are interested in, the applications they care about, and the very different challenges they face just in trying to do the science.
What are the biggest analytical challenges in trying to characterize food more comprehensively?
One of the biggest challenges in food analysis concerns the sheer complexity of food itself. My lab focuses primarily on small-molecule analysis – as well as elements – through what we call “ionomics,” where we look at broad panels of elements in food at the same time.
But food analysis is difficult because you cannot capture everything with one method. Naturally, that comes at a financial cost, and that is part of what has limited the field up to now. We can perform non-targeted small-molecule profiling using liquid chromatography, including reversed-phase and HILIC approaches, to capture a fairly broad range of compounds. Even then, those methods do not show us volatile compounds, proteins, or lipids.
One step we have introduced as part of our clean-up is to deliberately remove lipids, which helps generalize the method across matrices. However, this naturally means we do not capture lipids, fats, starches, and other components that we know are important. The PTFI is working on methods to address these different classes of molecules, but there are simply so many different types that matter. Even with high-performance analytical tools, you still have to focus on one class of molecules at a time.
Another major challenge is quantitation. A trade-off with using non-targeted approaches is that you do not get absolute quantitation. Historically, food composition work has been very targeted: you measure a small number of known compounds and get absolute concentrations.
That creates a disconnect and raises questions such as, “What do we do with relative quantitation? How do we use that information to drive discovery, inform decisions, make recommendations, or develop new hypotheses?” Part of the challenge is working across disciplinary boundaries and helping people think about food composition data differently from how it has traditionally been used.
Comprehensive analysis is also expensive. If you want to fully characterize a single food – from small molecules and elements to lipids, starches, proteins, amino acids, and macronutrients – it quickly becomes costly. In reality, we are moving toward a phased approach.
Right now, we have analyzed around 500 foods using all the methods currently available to us. We have tried to be as comprehensive as possible, although we are still not capturing everything. In the next phase, we hope to increase that by another 500 foods, bringing us to around 1,000 in total.
There is a balance here. On one hand, we want breadth: a wide range of foods characterized as comprehensively as possible to build a shared resource. On the other, we would like scientists to use these tools in focused ways depending on the specific questions they are trying to answer.
How do you see efforts like the PTFI changing how we think about or define food quality?
It is actually quite hard to imagine the initiative's ultimate potential. If you think about what the world looked like ten years ago, and how different it is now, it’s difficult to predict where this data will prove most valuable. Take tools such as Google Maps, for example. Early on, the idea was relatively simple: link images and location data so people could see places. Today, it has fundamentally changed how we live, travel, and navigate the world.
I think about this initiative in a similar way. We are generating this data, and if we do it right, people will come up with creative ways to use it – probably in ways we cannot even conceive. But I do think it has the potential to fundamentally change how we think about food, and how we think about the impacts of how we grow it.
There is also a broader cycle to consider. Agriculture is one of the biggest contributors to climate change, and the resulting environmental changes then affect the quality of the food we eat and how we grow it. Rockefeller is one of the main funders of this program, and the Rockefeller Foundation also played a major role in funding the Green Revolution. That helped enable large-scale agriculture and monocropping, producing vast amounts of food while also contributing to reduced biodiversity and climate pressures.
This initiative is an opportunity to shift that dialogue. Food sits at the center of climate change, human health, and sustainability. If we have more data, we can think more creatively about how food is grown, evaluated, and connected to those wider systems.
In your view, what impact could these developments have on society more broadly?
We are essentially creating a tool to enable and enhance future research. We are generating data, but we are also training people to use it.
One example comes from collaborators we have at UC Davis, who are thinking about how AI tools could make use of this kind of data. You could imagine developing a meal plan that is high in specific types of molecules, while also accounting for cultural preferences or restrictions around what different populations eat. You might learn that strawberries have a particular molecular profile of interest, but then work with a population that does not eat strawberries. With this kind of data, you could ask: what other foods have a similar molecular profile but would be more acceptable to that group?
In that way, we can take what we know about one food and translate it to others that meet similar nutritional or bioactive criteria. That is one exciting application of creating this kind of data resource.
We also have collaborators working on proteins and thinking about plant-based versus meat-based sources. The question there is how we identify plants that offer protein profiles most similar to animal-based sources, and how we can more intelligently design plant-forward meals that still meet protein requirements and nutritional needs.
Looking ahead five to 10 years, what needs to happen for this work to have real impact?
We are at an inflection point right now. We have established a strong foundation, and the next question is how we scale it.
Our next steps are focused on that question. What do we need to truly scale this effort? How can we automate more of what we are doing? How can we educate more people, such as by developing asynchronous training materials rather than relying on one-on-one calls? And how can we enable better access for our partners around the world to the reagents and tools they need?
At the same time, we are thinking analytically about how to expand the types of molecules we measure. One of our collaborating teams is working on volatile compounds, which are hugely important. No matter how healthy a food is, if it tastes bad, people are not going to eat it. Many of the flavor compounds that drive sensory perception are volatile compounds that we detect through smell.
We also have teams working on starches, fibers, sugars, and carotenoids – other bioactive compounds that are challenging to measure analytically, but just as important. The question is how we expand the scope of our analytical toolbox while, in parallel, determining how to scale the overall initiative.
We have built the foundation, but for this work to have real impact, we first have to learn how to scale it.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
