Calculates the effective number of tests performed in a GWAS, given the LD structure among the SNPs. SNPs that are highly correlated are not counted twice.
eff_nbr_tests(mat, cor.true = FALSE)
mat | matrix, either individual by SNPs (n x p) matrix, or correlation matrix between SNPs |
---|---|
cor.true | indicates if option mat is a correlation matrix or not |
zhc.correction
number of effective tests
Implemented from https://www.ncbi.nlm.nih.gov/pubmed/18271029
dat.sim <- data.frame(g = rnorm(100)) %>% dplyr::mutate(g2 = g * 2) eff_nbr_tests(diag(10), cor.true = TRUE) ## 10, because all are independent#> [1] 10eff_nbr_tests(dat.sim, cor.true = FALSE)#> [1] 1