Fig. 5
From: Prediction of BMI traits in the Chinese population based on the gut metagenome
The prediction of BMI by random forest regression analysis. (a) Model cross-validation error rate. (b) Species abundance regression model results with an R2 of 0.56 and an MAE of 2.09. (c) Species abundance and single-sample SNP density regression model results with an R2 of 0.62 and an MAE of 2.06. (d) Contig SNP regression model results with an R2 of 0.72 and an MAE of 1.56. The green line represents the species level, the yellow line accounts for both the species and gene level, and the orange‒red line pertains solely to the gene level