Extended Data Fig. 7: Feature molecules for SMI prediction models and SMI distribution in two cohorts. | Nature Aging

Extended Data Fig. 7: Feature molecules for SMI prediction models and SMI distribution in two cohorts.

From: Sarcosine decreases in sarcopenia and enhances muscle regeneration and adipose thermogenesis by activating anti-inflammatory macrophages

Extended Data Fig. 7

a, Changes in regression coefficients between characteristic variables and lambda values from Lasso machine learning for SMI prediction. b, Mean squared error changes between predicted and measured values versus lambda values. The left and right dotted lines represent the Lambda value (minimum mean square error) and the lambda value for the simplest model, respectively. c,d, Nuclear density plots SMI distribution in female (c) and male (d) participants across two cohorts. e,f, Violin plot showing the relative intensity of sarcosine in the normal, sarcopenic, and severe sarcopenic individuals in cohort1 (e) and cohort2 (f) (two-sided Mann‒Whitney U test; median ± quartiles; whiskers extend to minimum and maximum values; ‘N’ represents participant number). g-i, Scatter plot showing the Pearson correlation between sarcosine levels and SMI (g), HGS (h), and GS (i) in cohort1 (left) and cohort2 (right). Co means coefficients of Pearson correlation, and the p value was calculated by multiple linear regression adjusted with sex.

Back to article page