TGCA.decorrelate: Decorrelation Z-scores with estimated phenotypic correlations

In xiashen/TGCA: Total genetic contribution assessment

Description

The function re-weights GWAS Z-scores acorss multiple phenotypes by sample sizes and adjusts their phenotypic correlation before running TGCA.

Usage

TGCA.decorrelate(tstat, N, MAF, maf.cut = 5e-04, eigen.cut = 0.9)

Arguments

tstat	A data frame of GWAS Z-scores across K phenotypes of M SNPs.
N	A vector of the sample size of K phenotypes.
MAF	A vector of minor alle frequency of M SNPs.
maf.cut	A value defining the low-MAF cutoff, where the SNPs with lower MAF values will be used to estimate the phenotypic correlations. Default 5e-4.
eigen.cut	A value defining the proportion of information captured by the decorrelated eigenvectors. Default 0.9.

Value

A list of two elements, where z.decorrelated contains a matrix of decorrelated Z-scores corresponding to the M SNPs, and cor.pheno gives the estimated phenotypic correlation matrix.

Note

You can also use this fuction to re-weight GWAS Z-scores acorss multiple phenotypes by sample size and adjust their phenotypic correlation.

Author(s)

Ting Li, Xia Shen

References

Shen, X., Li, T., Ning, Z. (2020). Improved estimation of phenotypic correlations using summary association statistics. bioRxiv.

Examples

## Not run: 
data(tgca)
decor <- TGCA.decorrelate(tstat, MAF, N)
image(cor(decor$z.decorrelated))
image(decor$cor.pheno)

## End(Not run)

xiashen/TGCA documentation built on July 4, 2021, 3:50 p.m.