Skip to content

Latest commit

 

History

History
100 lines (70 loc) · 5.88 KB

README.md

File metadata and controls

100 lines (70 loc) · 5.88 KB

GTEx v8 GWAS analysis

Resources and analysis shared by "Widespread dose-dependent effects of RNA expression and splicing on complex diseases and traits".

Data

The results underlying these analyses can be found in zenodo.org:

Reproducible analysis for manuscript

The code for the manuscript's analyses is available here as R scripts with the following dependencies:

  • bigrquery
  • tidyverse
  • upsetR

To use BigQuery, it is helpful to install Google Cloud SDK and have read access to the required Google Cloud tables.

GTEx GWAS subgroup paper

The markdows can be built from an R session executing:

wflow_build("analysis/miscellaneous_statistics_2.Rmd")

The R scripts can be run from a bash session:

Rscript code/figures/figure_enloc_all_vs_eur.R

Manuscript material

Support scripts:

  • code/helpers folder contains miscellaneous R functions and definitions used throughtout the analyses.
    • code/helpers/_helpers_big_query_tables.R contains a centralized definition of Bigquery tables to be used by other scripts.
  • code/preprocess folder contains scripts that were ran once to setup auxiliary data.
    • code/preprocess/preprocess_gwas_regions.R counts detections per loci (independent LD regions) that will be used as inputs for other analyses.
    • code/preprocess/preprocess_setup_auxiliary_tables.R builds auxiliary tables in big query
    • code/preprocess/preprocess_mediation_analysis.R builds data for primary vs secondary concordance analysis. Can download the data from download_aux_data.sh.

Main Paper material:

  • code/figures/figure_enloc_all_vs_eur.R figure comparing ENLOC RCP when using all individuals vs using European only (Main Paper, suppl fig 24)

Companion paper material:

  • code/paper_material/tables.R Generates latex tables to be included in the paper. At the moment:
    • Supplementary Table S1: the list of 87 selected traits
    • Supplementary Table S2: expression and splicing models tally
  • code/paper_material/summaries_short.R: Numbers summarising numbers of genes, gene-tissue pairs, etc.
  • code/figures/gwas_imputation_deflation.R figure showing the deflation of GWAS' p-value distribution after imputation for 27 traits (Supplementary Figure S4)
  • code/figures/gwas_imputation_quality.R scatter plot of original vs imputed GWAS zscores (Supplementary Figure S3)
  • code/figures/predixcan_enloc_eqtl_sqtl.R plots summarising loci detection per emthod (Supplementary Figure S17, S18, S19, S20)
  • code/figures/models_gain.R plots a comparison of numbers of models between Elastic Net models and MASHR-based models
  • code/figures/proportions_bundle.R plots the proportion of enloc and s-predixcan detections (Supplementary Figure S14-A,B,C,D)
  • code/figures/upset_mashr_gwas_enloc_spredixcan.R generates upset plots underlying Supplementary Figure S15 (S-PrediXcan/enloc loci with MASHR models)
  • code/figures/upset_mashr_gwas_enloc_spredixcan.R generates upset plots underlying Supplementary Figure S16 (S-PrediXcan/enloc loci with Elastic Net models)
  • For more details about figure please go to code/

Markdowns:

  • analysis/miscellaneous_statistics_2.Rmd: This R markdown generates miscellaneous statistics and summaries from other methods results. i.e. This tallies summaries from S-PrediXcan results, enloc results; numbers of genomic loci with detections, etc.
  • analysis/gwas_enloc_predixcan_multixcan.Rmd: analyzes GWAS, S-PrediXcan, S-MultiXcan, enloc results and builds upset plots. This overlaps a bity with the previous markdown. Also Figure 5-e

A workflowr project.

Internal spreadsheet with GWAS-methods

gtex-gwas-paper-methods-spreadsheet

Deprecated

  • code/figures/FIG-DOSE-RESPONSE-CONCORDANCE-C.R scatter plot of beta prim vs. sec for Whole Blood, Europeans, rcp>0.1 (Figure 3-C)
  • code/figures/SFIG-CONCORDANCE-MEDIATING-EFFECTS-RANK-BY-EFFECT-SIZE.R scatter plot, residual plot, and p-value (Supplementary Figure S13)