Microbiome-based correction for random errors in nutrient profiles derived from self-reported dietary assessments.

Area of Science:

• Nutritional Epidemiology
• Computational Biology
• Bioinformatics

Background:

• Dietary intake assessment in large cohorts relies on self-reported data, prone to significant measurement errors.
• These errors lead to inaccuracies in nutrient profile calculations, limiting epidemiological study validity.
• Existing computational methods to correct these dietary assessment errors are scarce.

Purpose of the Study:

• To introduce a novel deep-learning approach, Microbiome-based nutrient profile corrector (METRIC).
• To leverage gut microbial composition for correcting random errors in self-reported dietary data.
• To evaluate METRIC's performance in improving nutrient profile accuracy.

Main Methods:

• Developed a deep-learning model, METRIC, integrating gut microbiome data.
• Applied METRIC to correct self-reported dietary data from 24-hour recalls and diet records.
• Validated METRIC using synthetic datasets and three real-world cohort datasets.

Main Results:

• METRIC effectively minimized simulated random errors in dietary assessments.
• The correction was particularly significant for nutrients metabolized by gut bacteria.
• Demonstrated robust performance across synthetic and real-world data.

Conclusions:

• METRIC shows strong potential for enhancing the accuracy of dietary intake data derived from self-report instruments.
• The microbiome-based approach offers a promising computational solution to a persistent challenge in nutritional epidemiology.
• Further validation is needed to assess METRIC's efficacy in correcting actual measurement errors.