xQTLImp: efficient and accurate xQTL summary statistics imputation

Motivation Quantitative trait locus (xQTL) analysis to multi-omic molecular features, such as gene transcription (eQTL), DNA methylation (mQTL) and histone modification (haQTL), is being widely used to decipher the effect of genomic variations on multi-level molecular activities. However, missing genotypes and limited effective sample size largely reduce the power of detecting significant xQTLs in single study or integrating multi-studies for meta-analysis. While existing hidden Markov models (HMMs) based imputation approaches require individual-level genotypes and molecular traits, there is still no available implementation suitable for imputation of xQTL summary statistics, which is becoming widely available and useful. Results We developed xQTLImp, a C++ software specially designed to efficiently impute xQTL summary statistics based on multivariate Gaussian approximation. Experiments on a single cell eQTL dataset demonstrated that considerable amount of novel significant eQTL associations can be rediscovered by xQTLImp. Availability Software is available at https://github.com/hitbc/xQTLimp. Contact ydwang@hit.edu.cn or jiajiepeng@nwpu.edu.cn Supplementary information Supplementary data are available online.