Obtain correlation coefficients and their variance-covariances

Usage

b2r(b, s, rho, n)

Arguments

b: the vector of linear regression coefficients.
s: the corresponding vector of standard errors.
rho: triangular array of between-SNP correlation.
n: the sample size.

Value

The returned value is a list containing:

r the vector of correlation coefficients.
V the variance-covariance matrix of correlations.

Details

This function converts linear regression coefficients of phenotype on single nucleotide polymorphisms (SNPs) into Pearson correlation coefficients with their variance-covariance matrix. It is useful as a preliminary step for meta-analyze SNP-trait associations at a given region. Between-SNP correlations (e.g., from HapMap) are required as auxiliary information.

References

Elston RC (1975). “On the correlation between correlations.” Biometrika, 62(1), 133-140. doi:10.1093/biomet/62.1.133 .

Becker BJ (2000). “Multivariate meta-analysis.” In Tinsley HE, Brown SD (eds.), Handbook of Applied Multivariate Statistics and Mathematical Modeling, chapter 17, 499-525. Academic Press, San Diego. ISBN 978-0126913606.

Casella G, Berger RL (2002). Statistical Inference, 2 edition. Duxbury. ISBN 978-0-534-24312-8.

Author

Jing Hua Zhao

Examples

if (FALSE) { # \dontrun{
n <- 10
r <- c(1,0.2,1,0.4,0.5,1)
b <- c(0.1,0.2,0.3)
s <- c(0.4,0.3,0.2)
bs <- b2r(b,s,r,n)
} # }