Because this matrix does not depend on the relationship between subjects in the study, unlike genotype imputation methods, DIST can be readily used to impute statistics for family association studies.
In this work, we aimed to minimize the MSE of imputed summary statistics in our objective function because imputation was one of our main focuses.
Imputed information statistics for rs78378222 and rs3585075 SNPs were 0.66 and 0.61, respectively.
To test for the joint effect of eQTL/functional SNPs known to affect the expression of a gene, JEPEG was designed to rely solely on the (univariate) measured and imputed summary statistics.
While there are much less demanding methods which avoid the genotype imputation step by directly imputing SNP statistics, e.g. Directly Imputing summary STatistics (DIST) proposed by our group, their implicit assumptions make them applicable only to ethnically homogeneous cohorts.
To directly impute summary statistics while (i) substantially reducing the computational burden and (ii) retaining imputation accuracy, we propose Direct Imputation of summary STatistics (DIST).
Because such an access is required by all genotype imputation methods, it is helpful to develop methods that impute summary statistics without going through the interim step of imputing genotypes.
DIST imputes the statistics at the unmeasured SNPs in a prediction window as a function of (i) the statistics at measured SNPs in a larger window, henceforth denoted as extended window, and (ii) the correlation matrix of both measured and unmeasured statistics, as estimated from a relevant reference dataset.
Within JEPEG software, DIST module is silently run to impute summary statistics for unmeasured functional SNPs.
Both methods can directly impute summary statistics (two-tailed Z-scores) for unmeasured SNPs from genome-wide association studies (GWASs) consisting of both family and independent cohorts.
To adequately analyze such cohorts, we propose DISTMIX, a very fast and novel method/software for directly imputing summary statistics of untyped makers from cosmopolitan cohorts without using subject-level genotype data.
