Prediction of host age and sex classification through gut microbes based on machine learning

The gut microbiome changes significantly during human aging. This work aims to establish the relationship between the gut microbiome and the age of the sample through regression algorithms in machine learning. The samples come from a completed study, which includes data on 1095 healthy Chinese people. We calculated the Spearman correlation coefficient between sample microbial abundance and age. The results showed that the abundance of Bifidobacterium was highly negatively correlated with the age of the sample, while Cellulosilyticum was positively correlated (and only significantly positively correlated in female samples). The regression model constructed by 19 kinds of microorganisms obtained after dimensionality reduction is used to predict the age of the host. At the same time, through the analysis of Spearman's coefficient, it is found that there are significant differences in the types and quantities of microorganisms in the intestines of male and female hosts. In this article, the classification and prediction of host sex by random forest classifier can achieve an accuracy of about 80%. •Based on the random forest classifier to predict the host gender, the accuracy rate is 80%.•Cellulosilyticum is the most important to the life expectancy of the model.•The fiber content of long-lived women is significantly higher than that of men.