It's 11:40 PM of an unexpectedly long day when Samanta and I are submitting the rebuttal of our paper: A reverse metabolic approach to weaning: in silico identification of immune-beneficial infant gut bacteria, mining their metabolism for prebiotic feeds and sourcing these feeds in the natural product space. We are in the middle of an extremely intense period, but we are committed to meet the deadline, as the results are part of the New Zealand trial "Nourish to Flourish": forty babies and their mothers took part in this study, which will follow the little ones until they turn one.

The project began a couple of years earlier, when COSBI and The Liggins Institute thought about a radically different way to approach nutrition studies, compared to traditional ones, and started a fruitful collaboration. Having identified the research goal, which is to improve infant immune health, we designed and developed an in silico Systems Nutrition strategy. The health benefit we look for with an improvement of the immune system is the reduction in the incidence of infections around weaning.

Human gut is host to the largest part of our immune system and this is not surprising when we think about our intestine not as an inner organ but as the sheltering conduct where the external environment transits through our bodies. Since we are born, and even before, gut and external microbes evolve and co-exist with the immune cells and this interaction, ultimately leading to the immune system “training”, is even well recognized in our everyday decisions based on a “gut feeling”. The immune system gets exposed to a variety of substances: dirt, allergens and environmental factors that help recognizing what is tolerable from what needs to be fought. This education is crucial to avoid unneeded over-responses or lacking defenses all along our lives. Gut bacteria ecosystem represents a huge variety of what we can commonly refer to as “good” and “bad” microorganisms that eventually need to come to a balance with our immune system.

Our research began identifying in silico, which means by advanced mining of public domain literature, infant gut microbes that could strengthen the immune system. Our starting point was the scientific literature where a search on ‘microbiome’ yielded more than 47,000 PubMed articles. Therefore our first need was to concentrate the efforts on those articles concerning the beneficial role of gut microbes on the infant immune system and infections prevention during weaning. We built specific queries to find only the relevant literature. Then we developed and applied natural language processing methods (NLP) in a multi-level text-mining pipeline to extract only the relevant sentences. We went from thousands of papers to hundreds of informative sentences. From these sentences, the gut microbes names were automatically recognized and extracted, resulting in 18 putative probiotics, selected for their contribution to the immune system development and their presence in the infant gut.

To promote and take care of these “good” bacteria we wanted to find out how we could feed their nutritional requirements: the so-called exogenous metabolites that represent their potential prebiotic food. This means that those metabolites cannot be produced by metabolic reactions, but they have to be provided from the environment, in this case the infant gastro-intestinal tract. To stimulate the growth of the health-beneficial gut bacteria (probiotics) the introduction of prebiotics into the diet is in fact a common procedure. Prebiotics are foods that can positively stimulate gut microbes harbored by the human colon. The nutrient profile of our candidate probiotic community was analyzed and, using a metabolic network topology approach, we identified the metabolites that are exogenously required by every microbe in the community.

Considering the exogenous compounds are both lacking and subject to bacteria competition, they are effectively limiting elements for the growth of our candidate microbes. So, we searched for these prebiotics in the form of complementary foods that could supply limiting nutrients to the community and positively modulate the infant gut composition. We mined specific databases to search for these essential metabolites into the natural food space. The obtained list of foods was manually screened to exclude clearly inadequate weaning aliments such as processed foods and alcoholic and we ranked them according to their efficiency in stimulating bacteria growth.

This unbiased approach led to identifying a range of natural foods satisfying our goal of an increased immune-beneficial gut microflora in humans, but the aliments had to be assessed by the nutritionists in our team with screening criteria that clearly included the adherence to pediatric recommendations. For example, while honey was the top hit, using it as a complementary food is controversial (in particular raw honey) or likewise for peanuts and meat which may come with age-specific recommendations. Among all the foods showing the highest amounts of essential vitamins and amino acids, the Kumara (New Zealand sweet potato) appeared a success as it is a practical choice for the formulation of infant complementary foods, besides a New Zealand product. Kumara powder was therefore proposed for the High-Value Nutrition National Science Challenge pilot clinical trials on infants aged 6–12 months, where each subject will receive a fixed quantity that could be mixed with the infant’s complementary foods.

The different steps of our pipeline required a very vast and varied expertise and involved different ‘players’ within our multidisciplinary and multinational team (microbiologists, bioinformatics and nutritionists). The overall research was scientifically and personally rewarding, and all the team is looking forward to the completion of the clinical trials that are adopting, while writing, the foods identified in the study to help validate this approach.