In GenuityScience/topr: Create Custom Plots for Viewing Genetic Association Results

knitr::opts_chunk$set(
  message = FALSE,
  collapse = TRUE,
  comment = "#>"
  )
library(topr)
library(magrittr)
set.seed(0)

The locuszoom function allows us to entend the functionality of regionplot when LD R2 values are available. One of the sample datasets, CD_UKBB also has a version with R2 values for these examples, R2_CD_UKBB.

head(R2_CD_UKBB)

locuszoom(R2_CD_UKBB)

Annotate the variants with vertical lines highlighting their positions on the plot:

locuszoom(
  R2_CD_UKBB, 
  annotate_with_vline=1e-8
)

Include a region size to show separate peaks:

locuszoom(
  R2_CD_UKBB,
  annotate_with_vline=1e-8,
  region_size=1e5
)

Obtaining LD R2 values for Locuszoom plot

Not always will the R2 values be already included in your results data.

Extract the snp of interest rs7713270, from the inbuilt CD_UKBB dataset.

lead_snp <- CD_UKBB %>%
  dplyr::filter(ID == "rs7713270") 
snp <- paste0(lead_snp$CHROM,":",lead_snp$POS)

Retrieve the R2 values with LDlinkR::LDproxy()

Run LDproxy to get variants correlated with rs7713270. Use the European (EUR) 1000 Genomes Project population and genome build GRCh38.

Note! To be able to run LDproxy, you have to register and get a token (see https://ldlink.nih.gov/?tab=apiaccess).

SNP <- LDlinkR::LDproxy(
  snp = snp,
  pop = "EUR",
  r2d = "r2",
  token = "NULL",
  genome_build = paste0("grch", "38")
)
LD.link <- SNP %>% 
  dplyr::mutate(CHROM = stringr::str_extract(Coord, "chr[^:]*")) %>%
  dplyr::mutate(POS = as.numeric(stringr::str_extract(Coord, "(?<=:)[0-9]*"))) %>%  
  dplyr::select(CHROM, POS, R2)

Plot with the locuszoom function

Join the output from LDproxy with the CD_UKBB dataset to get the p-values for plotting.

snps.ld <- dplyr::inner_join(CD_UKBB, LD.link, by = c("CHROM", "POS")) 
locuszoom(snps.ld)

GenuityScience/topr documentation built on Oct. 20, 2024, 10:23 p.m.