Map GWAS Hits to LD Blocks (Post-GWAS QTL Region Definition)

Maps significant GWAS markers onto LD blocks to define QTL regions. Blocks with significant markers from multiple traits are flagged pleiotropic. Implements the approach of Tong et al. (2024).

Usage

define_qtl_regions(
  gwas_results,
  blocks,
  snp_info,
  p_threshold = 5e-08,
  trait_col = "trait",
  min_snps = 3L,
  ld_decay = NULL,
  verbose = TRUE
)

Arguments

gwas_results

Data frame with columns SNP (or Marker), CHR, POS. Optional columns: P (p-value), BETA (additive effect estimate), trait.

blocks

LD block data frame from run_Big_LD_all_chr.

snp_info

Full SNP metadata data frame.

p_threshold

Significance threshold. Default 5e-8. NULL uses all markers regardless of p-value.

trait_col

Trait column name in gwas_results. Default "trait".

min_snps

Minimum SNPs per block. Default 3L.

ld_decay

Optional LDxBlocks_decay object from compute_ld_decay, or a data frame with columns CHR and decay_dist_bp. When supplied, candidate gene windows are extended by the chromosome-specific decay distance on both sides of each lead SNP position, adding columns candidate_region_start, candidate_region_end, and candidate_region_size_kb to the output. Default NULL.

verbose

Logical. Print progress messages. Default TRUE.

Value

Data frame with one row per block containing significant markers: block_id, CHR, start_bp, end_bp, n_snps_block, n_sig_markers, lead_marker, lead_p, lead_beta (if BETA supplied), sig_markers, sig_betas (if BETA supplied), traits, n_traits, pleiotropic.

Block effect estimation

When multiple SNPs within a block are GWAS-significant, their marginal BETA values are correlated (due to LD) and cannot be summed directly. LDxBlocks returns three complementary columns when BETA is present in gwas_results:

lead_beta

BETA of the lead (lowest-p) SNP. The simplest proxy for the block effect – assumes the lead SNP fully tags the causal variant. Underestimates the block effect when multiple independent signals exist.

sig_markers

Semicolon-separated IDs of all significant SNPs in the block. Pass these to a conditional/joint analysis tool (e.g. COJO, SuSiE, finemap) to obtain independent within-block effects.

sig_betas

Semicolon-separated marginal BETA values for all significant SNPs (same order as sig_markers). Their absolute values are an upper bound on the true block effect because they include LD-induced inflation; use lead_beta or joint analysis for calibrated estimates.

For the biologically correct block-level effect, fit a haplotype model: build_haplotype_feature_matrix() produces the 0/1/2 dosage columns whose regression coefficients (alpha_h) are the true per-allele effects.

References

Tong J, Tarekegn ZT, Jambuthenne D, Alahmad S, Periyannan S, Hickey L, Dinglasan E, Hayes B (2024). Stacking beneficial haplotypes from the Vavilov wheat collection to accelerate breeding for multiple disease resistance. Theoretical and Applied Genetics 137:274. doi:10.1007/s00122-024-04784-w

Yang J et al. (2012). Conditional and joint multiple-SNP analysis of GWAS summary statistics identifies additional variants influencing complex traits. Nature Genetics 44(4):369-375. doi:10.1038/ng.2213