DISSCO: Direct Imputation of Summary Statistics allowing COvariates

DISSCO is a JAVA application for direct imputation of summary statistics allowing for covariates.

• If individual level genotypes are available, we highly recommend performing imputation at genotype level, using standard software like MaCH/MaCH-Admix/Minimac, IMPUTE/IMPUTE2 or BEAGLE.
• However, sometimes, individual level genotypes are NOT readily available, for example, in meta-analysis. Under such scenarios, association summary statistics (ASS) such as odds ratio, effect size estimate, Z score, Chi-square statistic, P-value, are usually readily available. DISSCO imputes ASS (particularly Z-score) of untyped markers using ASS of typed markers.
• Compared with DIST and ImpG-Summary/LD, DISSCO allows for the adjustment of covariates such as age, gender, ethnicity, top PCs and other confounders, which are almost inevitable in real genetic studies.

The current version is a pre-release. For details, please refer to our tutorial. Comments and suggestions are welcome, please e-mail Zheng Xu at xuzheng@email.unc.edu or Yun Li at yunli@med.unc.edu.