# GWAS summary statistics imputation, integration with PrediXcan MASHR-M The file `sample_data.tar` contains all necessary files to perform imputation of GWAS summary statistics to the GTEx v8 QTL data set. It includes 1000 Genomes individuals' genotypes as reference panel. The `.tar` archive, upon uncompression, contains the following: ``` data/ ├── eur_ld.bed.gz ├── gtex_v8_eur_filtered_maf0.01_monoallelic_variants.txt.gz ├── coordinate_map ├── gwas ├── liftover ├── models │ ├── eqtl │ │ └── mashr │ └── sqtl │ └── mashr └── reference_panel_1000G ``` `data/eur_ld.bed.gz` contains definitions of approximately independent LD-regions in hg38 (Berisa-Pickrell regions, lifted over) `data/gtex_v8_eur_filtered_maf0.01_monoallelic_variants.txt.gz` is a snp annotation file, listing all GTEx v8 variants with MAF>0.01 in europeans. `data/coordinate_map` contains precomputed mapping tables that MetaXcan tools can use to convert GWAS' genomic coordinates in GWAS between genome assemblies. `data/gwas` contains a sample GWAS file for the purposes of a tutorial (data obtained from Nikpay et al (Nat Gen 2016) https://www.ncbi.nlm.nih.gov/pubmed/26343387 `data/liftover` contains Liftover chains to map coordinates between human genome assemblies (used by full harmonization tools) `data/models` contains PrediXcan MASHR-M models, and cross-tissue S-MultiXcan LD compilation, from eQTL and sQTL. `data/reference_panel_1000G` contains 1000G hg38 genotypes, in parquet format, to be used by imputation tools.

{"references": ["Nikpay et al (Nat Gen 2016) \"A comprehensive 1000 Genomes\u2013based genome-wide association meta-analysis of coronary artery disease\" doi 10.1038/ng.3396", "Barbeira et al (Biorxiv 2019) \"Widespread dose-dependent effects of RNA expression and splicing on complex diseases and traits\" doi 10.1101/814350"]}