Skip to contents

Decompose Per-SNP Effects into Per-Haplotype-Allele Effect Table

Source: R/haplotype_analysis.R
decompose_block_effects.Rd

Aggregates the per-SNP additive effects (from backsolve_snp_effects) into a per-allele effect for each LD block: the effect of carrying allele h is the sum of SNP effects weighted by the allele's dosage at each SNP position. Returns an interpretable table of "which allele of which block is worth how many units of the trait?"

Usage

decompose_block_effects(
  haplotypes,
  snp_info,
  blocks,
  snp_effects,
  min_freq = 0.02,
  missing_string = "."
)

Arguments

haplotypes

Named list from extract_haplotypes.

snp_info

Data frame with SNP, CHR, POS.

blocks

LD block table from run_Big_LD_all_chr.

snp_effects

Named numeric vector of per-SNP additive effects, as returned by backsolve_snp_effects.

min_freq

Numeric. Alleles below this frequency are excluded. Default 0.02.

missing_string

Character. Missing haplotype placeholder. Default ".".

Value

Data frame with one row per allele per block, sorted by CHR, start_bp, effect_rank.

block_id, CHR, start_bp, end_bp

Block coordinates.

allele

Haplotype allele string.

frequency

Allele frequency in the panel.

allele_effect

Sum of SNP effects weighted by allele dosage.

effect_rank

Rank within block; 1 = most positive effect.

n_snps_block

Number of SNPs in this block.

See also

backsolve_snp_effects, rank_haplotype_blocks

Examples

# \donttest{
data(ldx_geno, ldx_snp_info, ldx_blocks, ldx_blues, package = "LDxBlocks")
haps    <- extract_haplotypes(ldx_geno, ldx_snp_info, ldx_blocks)
res     <- run_haplotype_prediction(ldx_geno, ldx_snp_info, ldx_blocks,
                                     blues = ldx_blues, id_col = "id",
                                     blue_col = "YLD", verbose = FALSE)
# Single-trait: res$snp_effects is the named numeric vector directly
snp_fx  <- res$snp_effects
allele_tbl <- decompose_block_effects(haps, ldx_snp_info, ldx_blocks,
                                       snp_effects = snp_fx)
head(allele_tbl[order(-allele_tbl$allele_effect), ])
#>                                          block_id CHR start_bp end_bp
#> 222222222222222222222022202222 block_2_1000_30023   2     1000  30023
#> 122121212112211211122011201112 block_2_1000_30023   2     1000  30023
#> 111122112222221222211122211112 block_2_1000_30023   2     1000  30023
#> 022020202002200200022000200002 block_2_1000_30023   2     1000  30023
#> 011021102112210211111111210002 block_2_1000_30023   2     1000  30023
#> 000022002222220222200222220002 block_2_1000_30023   2     1000  30023
#>                                                        allele frequency
#> 222222222222222222222022202222 222222222222222222222022202222    0.0250
#> 122121212112211211122011201112 122121212112211211122011201112    0.0417
#> 111122112222221222211122211112 111122112222221222211122211112    0.0917
#> 022020202002200200022000200002 022020202002200200022000200002    0.0333
#> 011021102112210211111111210002 011021102112210211111111210002    0.0917
#> 000022002222220222200222220002 000022002222220222200222220002    0.1000
#>                                allele_effect n_snps_block effect_rank
#> 222222222222222222222022202222      1.487853           30           1
#> 122121212112211211122011201112      1.325151           30           2
#> 111122112222221222211122211112      1.265583           30           3
#> 022020202002200200022000200002      1.162448           30           4
#> 011021102112210211111111210002      1.102880           30           5
#> 000022002222220222200222220002      1.043313           30           6
# }