Evaluation of gut microbiota predictive potential associated with phenotypic characteristics to identify multifactorial diseases.
Danielle Cristina FonsecaIlanna Marques Gomes da RochaBianca Depieri BalmantLeticia CalladoAna Paula Aguiar PrudêncioJuliana Tepedino Martins AlvesRaquel Susana TorrinhasGabriel da Rocha FernandesDan Linetzky WaitzbergPublished in: Gut microbes (2024)
Our findings underscore that the selection of bacterial taxa based solely on differences in relative abundance between groups is insufficient to serve as clinical markers. Machine learning techniques are essential for mitigating the considerable variability observed within gut microbiota. In our study, the use of microbial taxa alone exhibited limited predictive power for health outcomes, while the integration of phenotypic variables into predictive models substantially enhanced their predictive capabilities.