Metabolic dysfunction-associated steatotic liver disease (MASLD) is a common and complex condition, affecting an estimated 25–30% of the global population. MASLD is considered a part of metabolic syndrome and strongly linked to factors such as obesity and type 2 diabetes. However, there is currently a lack of reliable non-invasive biomarkers for early detection of MASLD, leading to an underdiagnosis of the condition and poor patient outcomes. Our research project aims to address this by identifying multi-omic, including metagenomic, biomarkers of MASLD in the general population.
As a primary study in this project, we are conducting a multi-omics analysis integrating the fecal microbiome and host systemic responses in a Polish population-based cohort of 400 participants aged 35–65. Shotgun metagenomic data from stool samples will be processed using custom and established tools (MetaPhlAn, StrainPhlan, HUMAnN) to identify microbial taxa and functional pathways associated with MASLD across different disease presentations. These will be analysed alongside metabolic risk factors and stratified by key demographic variables. Machine learning approaches (e.g. random forest classifiers) will be used to identify features that distinguish the healthy individuals and different MASLD subgroups, such as "lean MASLD".
To validate and further contextualise these findings, we aim to use the Dutch Microbiome Project (DMP) dataset as an external population-scale reference. Although MASLD diagnosis data is not available in DMP, its broader demographic and geographic coverage, paired with age, sex, and BMI metadata, will help assess the distribution of the candidate metagenomic features relevant to MASLD in the wider general population (as our cohort is restricted to a single municipality). We plan to evaluate whether the observed prevalence and distribution of these microbial signatures in DMP aligns with the expected epidemiological trends of MASLD across the basic demographic (i.e., age and BMI- determined) groups, thus supporting the robustness and generalisability of our results.
Furthermore, to support the efficiency and feasibility of our research timeline, we plan to use the DMP dataset to optimise and benchmark our bioinformatic pipelines before the metagenomics data from our own cohort becomes available. The DMP cohort will be valuable for this purpose given the shared European context and broadly comparable environmental, lifestyle and socio-economic patterns between the Netherlands and Poland, as well as technological similarities in the shotgun data generation. Moreover, testing out bioinformatic methods across an established dataset will help ensure the reproducibility of our analysis.
The project will be carried out over 12 months, beginning with pipeline development and optimisation (months 1–4), followed by cross-cohort microbiome and epidemiological distribution analysis (months 5–8), and final synthesis of results and manuscript preparation (months 9–12). The inclusion of the DMP dataset in this study will help understand the epidemiological spread of MASLD-related microbial features across populations, optimize metagenomics data processing tools, and ultimately support the development of non-invasive diagnostics for metabolic liver disease.