Detect mismatch between GWAS zscores and LD reference using SuSiE RSS

Last updated: 2023-11-14

Checks: 7 0

Knit directory: multigroup_ctwas_analysis/

This reproducible R Markdown analysis was created with workflowr (version 1.7.0). The Checks tab describes the reproducibility checks that were applied when the results were created. The Past versions tab lists the development history.

R Markdown file: up-to-date

Great! Since the R Markdown file has been committed to the Git repository, you know the exact version of the code that produced these results.

Environment: empty

Great job! The global environment was empty. Objects defined in the global environment can affect the analysis in your R Markdown file in unknown ways. For reproduciblity it’s best to always run the code in an empty environment.

Seed: set.seed(20231112)

The command set.seed(20231112) was run prior to running the code in the R Markdown file. Setting a seed ensures that any results that rely on randomness, e.g. subsampling or permutations, are reproducible.

Session information: recorded

Great job! Recording the operating system, R version, and package versions is critical for reproducibility.

Cache: none

Nice! There were no cached chunks for this analysis, so you can be confident that you successfully produced the results during this run.

File paths: relative

Great job! Using relative paths to the files within your workflowr project makes it easier to run your code on other machines.

Repository version: df9cc88

Great! You are using Git for version control. Tracking code development and connecting the code version to the results is critical for reproducibility.

The results in this page were generated with repository version df9cc88. See the Past versions tab to see a history of the changes made to the R Markdown and HTML files.

Note that you need to be careful to ensure that all relevant files for the analysis have been committed to Git prior to generating the results (you can use wflow_publish or wflow_git_commit). workflowr only checks the R Markdown file, but you know if there are other scripts or data files that it depends on. Below is the status of the Git repository when the results were generated:


Ignored files:
    Ignored:    .Rproj.user/

Unstaged changes:
    Modified:   analysis/index.Rmd

Note that any generated files, e.g. HTML, png, CSS, etc., are not included in this status report because it is ok for generated content to have uncommitted changes.

These are the previous versions of the repository in which changes were made to the R Markdown (analysis/detect_LD_mismatch_gwas_UKBBref_susie_rss.Rmd) and HTML (docs/detect_LD_mismatch_gwas_UKBBref_susie_rss.html) files. If you’ve configured a remote Git repository (see ?wflow_git_remote), click on the hyperlinks in the table below to view the files as they were in that past version.

File	Version	Author	Date	Message
Rmd	df9cc88	kevinlkx	2023-11-14	wflow_rename(“analysis/detect_LD_mismatch_gwas_UKBBref.Rmd”,
html	df9cc88	kevinlkx	2023-11-14	wflow_rename(“analysis/detect_LD_mismatch_gwas_UKBBref.Rmd”,

Load packages and functions

library(ctwas)
library(susieR)
library(tidyverse)
library(GenomicRanges)
library(rtracklayer)

# Load UKBB reference LD matrix and SNP info
load_UKBB_R_snp_info <- function(region_df, ld_R_dir, filestem = "ukb_b38_0.1"){
  filename <- sprintf("%s_chr%s.R_snp.%d_%d", filestem,
                      gsub("chr", "", region_df$chr), region_df$start, region_df$stop)
  print(filename)
  if(!file.exists(file.path(ld_R_dir, paste0(filename, ".RDS"))) || !file.exists(file.path(ld_R_dir, paste0(filename, ".Rvar")))){
    stop("LD Reference files not exist!")
  }
  R_snp <- readRDS(file.path(ld_R_dir, paste0(filename, ".RDS")))
  R_snp_info <- read.table(file.path(ld_R_dir, paste0(filename, ".Rvar")), header=TRUE)
  return(list(R_snp = R_snp, R_snp_info = R_snp_info))
}

# Match GWAS sumstats with LD reference files. Only keep variants included in LD reference.
match_gwas_R_snp <- function(sumstats, R, snp_info){
  sumstats <- sumstats[sumstats$id %in% snp_info$id,]
  R_snp_index <- na.omit(match(sumstats$id, snp_info$id))
  sumstats$R_snp_index <- R_snp_index
  R <- R[R_snp_index, R_snp_index]
  stopifnot(nrow(sumstats) == nrow(R))
  return(list(sumstats = sumstats, R = R))
}

liftOver_hg19ToHg38 <- function(gr){
  seqlevelsStyle(gr) <- "UCSC"
  ch <- import.chain("~/softwares/liftOver/hg19ToHg38.over.chain")
  gr <- unlist(liftOver(gr, ch))
  genome(gr) <- "hg38"
  return(gr)
}

Load LD Regions (ldetect blocks)

regions <- system.file("extdata/ldetect", "EUR.b38.bed", package = "ctwas")
regions_df <- read.table(regions, header = T)
regions_df <- regions_df %>% dplyr::arrange(chr, start, stop) %>% dplyr::mutate(locus = 1:nrow(regions_df))

LDL

trait <- "LDL"
set.seed(123)
outdir <- paste0("/project2/xinhe/kevinluo/cTWAS/multigroup/ld_mismatch_filtering/", trait)
if(!dir.exists(outdir)) dir.create(outdir)

GWAS data

# read the data
sumstats <- VariantAnnotation::readVcf("/project2/xinhe/shared_data/multigroup_ctwas/gwas/ukb-d-30780_irnt.vcf.gz")
sumstats <- gwasvcf::vcf_to_tibble(sumstats)

# compute the z-scores
sumstats$Z <- sumstats$ES/sumstats$SE

# drop multiallelic variants (id not unique)
sumstats <- sumstats[!(sumstats$rsid %in% sumstats$rsid[duplicated(sumstats$rsid)]),]

keep_columns <- c("seqnames", "start", "rsid", "REF", "ALT", "ES", "SE", "Z", "LP")
sumstats <- sumstats[,keep_columns] %>% 
  dplyr::rename(chr = seqnames, pos = start, id = rsid, A1 = ALT, A2 = REF, z = Z, pval = LP)
saveRDS(sumstats, file="/project2/xinhe/shared_data/multigroup_ctwas/gwas/gwas_processed/LDL-ukb-d-30780_irnt.sumstats.RDS")

# subset the columns and format the column names
z_snp <- as.data.frame(sumstats[,c("id", "A1", "A2", "z")])

# save the formatted z-scores
saveRDS(z_snp, file="/project2/xinhe/shared_data/multigroup_ctwas/gwas/gwas_processed/LDL-ukb-d-30780_irnt.z_snp.RDS")

sumstats_file <- "/project2/xinhe/shared_data/multigroup_ctwas/gwas/gwas_processed/LDL-ukb-d-30780_irnt.sumstats.RDS"
print(sumstats_file)
sumstats <- readRDS(sumstats_file)
z_snp <- as.data.frame(sumstats[,c("id", "A1", "A2", "z")])
gwas_n <- 343621

Assign GWAS SNPs to LD Regions

sumstats <- makeGRangesFromDataFrame(sumstats, start.field = "pos", end.field = "pos", keep.extra.columns = T)
sumstats_hg38 <- liftOver_hg19ToHg38(sumstats) %>% as.data.frame() %>%
  dplyr::rename(chr = seqnames, pos = start, snp = id) %>%
  dplyr::select(chr, pos, snp, A1, A2, z, pval)
sumstats_hg38 <- mapgen::assign_snp_locus(sumstats_hg38, regions_df)

Select locus

sig.loci <- sumstats_hg38 %>% dplyr::filter(pval > -log10(5e-8)) %>%
  dplyr::arrange(-pval) %>%
  dplyr::pull(locus) %>% unique()
saveRDS(sig.loci,  file.path(outdir, paste0(trait, ".sig.loci.rds")))

rm(sumstats)
rm(sumstats_hg38)

LD mismatch diagnosis

sig.loci <- readRDS(file.path(outdir, paste0(trait, ".sig.loci.rds")))
cat(length(sig.loci), "significant loci.\n")

n.loci <- length(sig.loci)
condz_table <- readRDS(file.path(outdir, paste0(trait, ".lambda.condz.", n.loci, "loci.rds")))
condz_table <- condz_table %>% dplyr::rename(flip_snps = problematic_snps)
condz_table$n_flip_snps <- 0
condz_table$n_pval_snps <- 0

condz_dist_all <- readRDS(file.path(outdir, paste0(trait, ".condz.dist.", n.loci, "loci.rds")))

for(locus in condz_table$locus){
  condz_dist <- condz_dist_all[[as.character(locus)]]
  condz_dist$pval_diff <- pchisq(condz_dist$z_std_diff^2, df = 1, lower.tail=FALSE)
  
  detected_flip_index <- which(condz_dist$logLR > 2 & abs(condz_dist$z) > 2)
  detected_pval_diff_index <- which(condz_dist$pval_diff < 5e-8)

  condz_table[condz_table$locus == locus, ]$n_flip_snps <- length(detected_flip_index)
  condz_table[condz_table$locus == locus, ]$n_pval_snps <- length(detected_pval_diff_index)
}

sprintf("%d out of %d (%.2f %%) loci with problematic variants", length(which(condz_table$n_pval_snps > 0)), nrow(condz_table),
        length(which(condz_table$n_pval_snps > 0))/nrow(condz_table)*100)

sprintf("%d out of %d (%.2f %%) loci with flipped variants", length(which(condz_table$flip_snps != "")), nrow(condz_table),
        length(which(condz_table$flip_snps != ""))/nrow(condz_table)*100)

ggplot(condz_table, aes(x = n_flip_snps)) + 
  geom_histogram(bins = 30) +
  labs(x = "Number of flipped variants (logLR > 2 & abs(z) < 2)", y = "Number of loci") +
  theme_bw()

Version	Author	Date
df9cc88	kevinlkx	2023-11-14


ggplot(condz_table, aes(x = n_pval_snps)) + 
  geom_histogram(binwidth = 1) + 
  labs(x = "Number of problematic variants (p < 5e-8)", y = "Number of loci") +
  theme_bw()

Version	Author	Date
df9cc88	kevinlkx	2023-11-14


condz_table %>% dplyr::mutate(lambda = round(lambda, 4)) %>% 
  dplyr::select(locus, lambda, n_flip_snps, n_pval_snps) %>%
  DT::datatable(options = list(order = list(list(3, 'desc'), list(4, 'desc'))))
# 169 significant loci.
# [1] "147 out of 169 (86.98 %) loci with problematic variants"
# [1] "2 out of 169 (1.18 %) loci with flipped variants"

AFib

trait <- "AFib"
outdir <- "/project2/xinhe/kevinluo/cTWAS/multigroup/ld_mismatch_filtering/AFib"
if(!dir.exists(outdir)) dir.create(outdir)