Skip to main content

Table 2 Mediation effect size estimated using the Framingham Heart Study data.

From: Estimation of total mediation effect for high-dimensional omics mediators

Outcome

n

\({{\hat{p}}}\)

\(R^2_{Y,X}\)

\(R^2_{Mediated}\)

SOS

ab\(^{1}\)

prop

ratio

\(\nu \) \(^{1}\)

FVC\(^{2}\)

1378

0

0.207

0

0

0

0

0

0

 

(0, 0)

(0.153, 0.265)

(0, 0)

(0, 0)

(0, 0)

(0, 0)

(0, 0)

(0, 0)

Systolic BP

1711

207

0.069

0.026

0.381

0.002

0.008

0.008

4.1e−6

 

(146, 224)

(0.035, 0.111)

(− 0.003, 0.066)

(− 0.085, 0.771)

(− 0.04, 0.03)

(− 0.17, 0.14)

(− 0.14, 0.16)

(1.1e−6, 1.8e−3)

  1. 95% CI is within the parentheses based on percentiles of 500 bootstrap samples; \({{\hat{p}}}\) is the number of genes in estimation; n is the sample size for each trait; A mixed model is used to estimate the quantities, including \(R^2\)’s, ab (the product measure), prop (the proportion measure), ratio, and the \(\nu \) measure for multiple mediators
  2. 1ab and \(\nu \) were calculated based on standardized residuals with SD = 1
  3. 2Lasso and FDR methods were also applied on FVC, by which none of the gene was selected