SummaryOver the last decades, the incidence of colorectal cancer (CRC) in young adults has increased at an alarming rate. Global Westernization of diets, synthetic food dyes, physical inactivity and sedentary behavior, have been identified as key risk factors. The gut microbiota is at the intersection of early onset CRC and those risk factors and can be seen as a dynamic living sensor of any change in the human body. Therefore, leveraging recent advances in next generation sequencing can lead to novel microbiome-based prophylactics, diagnostics and therapeutics against CRC in young adults (overall objective of the consortium). Accumulated evidence suggests that development of CRC is the result of damage years earlier and that early life exposures contribute to the genesis of the disease later in life. The goal of the project is to establish preventive measures to reduce the risk of CRC development in young age. To filter out the main suspects affecting CRC, we must consider that development of CRC in young age is increasing since the 80’s where the “Westernized diet and lifestyle” was adopted. It is a global phenomenon, especially in developed countries, and it seems to be associated with gut microbiome dysbiosis. As such, focus will be given to early life exposomal elements that have microbiome-modifying properties, and which allow either the establishment of pathogenic microbes or deplete the gut from beneficial ones. We will perform metagenomics sequencing of stool samples from around 550 healthy young adults in total to associate in unprecedent detail the nutritional behavior, lifestyle choices and medication of these subjects with microbiome composition. With this knowledge in hand, we will subsequently apply microbial source tracking algorithms to reveal dietary and lifestyle patterns that promote or inhibit the establishment of the CRC microbiome signatures (established in other subprojects of PerMiCCion). A major breakthrough is expected by defining age-windows of vulnerability (for disease risk) and opportunity (for health promotion) that can be leveraged for prevention. Lastly, our dissemination and exploitation plan targets both the population level - Digestive Cancers - to communicate findings in relevant policy-making bodies, and individual level - DNA NutriCoach - to design nutritional and lifestyle recommendations based on the metagenomic profile of each individual. In our study we plan to infer generalized sources of heterogeneity in the human gut microbiota of young adults and identify lifestyle characteristics, presence of other diseases and medications that have the greatest impact in shaping the microbial community composition and activity in young adults. (1) Paired sample studies for ML analysis will be constructed to assess the strength of association between each host variable and the composition of gut microbiota (bacteria and fungi). Data about age, sex, lifestyle, disease status and medications collected from all participants allow simple constructions of case-controls; (2) Metagenomics sequencing (shotgun metagenomics for the microbiome and ITS for the mycobiome) will be applied for 550 subjects (<40 years old, 50:50 ratio of males and females) included in the GNC cohort and a table of the gut taxonomic composition (bacteria and fungi) and functional activity (bacteria) will be generated. Binary cohort ML analysis will be performed to infer the ability of gut microbiota to discriminate between the cases and controls for each variable. The accuracy and discriminatory power of the models will be assessed using the AUC. Cohorts for all lifestyle variables in the questionnaires that have >25 samples will be constructed. We expect to achieve the most comprehensive and detailed associations between lifestyle, medication, and gut bacteria and fungi in young adults.

Keywords colon-cancer gut machine-learning microbiome mycobiome

InstitutionsUniversitätsklinikum Freiburg, Institut für Prävention und Tumorepidemiologie, Leibniz-HKI, Universitaet Freiburg