Comparison of methods for transcriptome imputation through application to two common complex diseases

• DOI: 10.1038/s41431-018-0176-5
• Publication Date: 2018-07-05T11:17:45Z
• Authors: James J. Fryett; Jamie Inshaw; Andrew P. Morris; Heather J. Cordell
• Keywords: Genome-wide Association Study; Imputation (statistics); Genetic Association

Transcriptome imputation has become a popular method for integrating genotype data with publicly available expression to investigate the potentially causal role of genes in complex traits. Here, we compare three approaches (PrediXcan, MetaXcan and FUSION) via application genome-wide association study (GWAS) Crohn's disease type 1 diabetes from Wellcome Trust Case Control Consortium. We investigate: (i) how results each approach other those standard GWAS analysis; (ii) variants models used by prediction tools previously reported as eQTLs. find that all produce highly correlated when applied same data, although subset genes, mostly major histocompatibility complex, strongly disagree. also observe most associations detected these methods occur near known risk loci. Application transcriptome summary statistics meta-analyses detects 53 significant expression—Crohn's 154 expression—type associations, providing insight into biology underlying diseases. conclude while current implementations typically detect fewer than GWAS, they nonetheless provide an interesting way interpreting signals identify genes.