Predictive modelling for analyzing characteristics of gut microbiota in different age groups of healthy population

  Objective  To explore the characteristics of gut microbiota in different age groups by establishing a predictive model based on 16S rRNA gene sequencing data in healthy population.

  Methods  The featured OTUs were screened based on 923 samples in 2016 from SRA database, and a predictive model of different groups was established by random forest model to analyze the differential characteristics and function of gut microbiota among different age groups and calculate the correlations between the featured OTUs and the differential pathways.

  Results  The accuracy of random forest model was 72.9% based on 278 featured OTUs associated with age. There were 249 differential pathways among different age groups, mainly in xenobiotics biodegradation, metabolism, lipid metabolism and amino acid metabolism. According to coliform group and function, pupil and mid-age were grouped together, and youth, elderly and centenarian were classified into the same cluster.

  Conclusion  The predictive model and clustering results show that the characteristics of gut microbiota in the elderly, centenarian and youth are similar.