Microbiome-based correction of nutrient profiles derived from self-reported dietary assessments.

Since dietary intake is hard to directly measure in large-scale cohort studies, we often rely on self-reported instruments (e.g., food frequency questionnaires, diet recall surveys, and diet diary methods) developed in nutritional epidemiology. Those self-reported instruments are prone to measurement errors. The measurement errors eventually lead to inaccuracies in the calculation of nutrient profiles. Currently, there is a lack of computational methods to address this problem. To fill the gap, we introduce a deep-learning approach --- M icrobiom e -based nu t rient p r of i le c orrector (METRIC), which leverages gut microbial compositions to correct the errors in nutrient profiles due to measurement errors in self-reported dietary assessments. We demonstrate the excellent performance of METRIC in minimizing the simulated random errors in both synthetic and three real-world datasets. METRIC has the potential to significantly improve the accuracy of self-reported dietary assessments and hence facilitate the research of nutritional epidemiology.