Split unphased genotype matrix by ancestry for K populations

Generalises split_diploid to an arbitrary number of ancestries. For each variant, population allele proportions \(p_1, \ldots, p_K\) (summing to 1) are estimated from samples with unambiguous ancestry (pure-ancestry homozygotes and heterozygotes, plus mixed hom-alts which each contribute one allele to each parent pool). Ambiguous heterozygotes in each mixed pair are then split by the conditional pairwise ratio \(p_i / (p_i + p_j)\). When no unambiguous alt alleles exist for a pair (singleton edge case), the split defaults to 0.5 / 0.5.

split_diploid_multi(
  gt_genotype,
  ancestry,
  pure_codes,
  mixed_codes,
  gla = NULL,
  arm_id = NULL
)

Arguments

gt_genotype

Integer matrix (variants x samples) of genotypes (0/1/2).

ancestry

Integer matrix (variants x samples) of ancestry codes, same dimensions as gt_genotype.

pure_codes

Named integer vector mapping each population label to its pure-ancestry diploid code. Example: c(AFR = 3L, EUR = 1L, NAT = 4L).

mixed_codes

Data frame with three columns:

code: Integer ancestry code for this mixed pair.
pop1: Character name of the first parent population (must be a name in pure_codes).
pop2: Character name of the second parent population (must be a name in pure_codes).

One row per mixed-ancestry diploid type. Example for 3 populations: data.frame(code=c(2L,5L,6L), pop1=c("AFR","AFR","EUR"), pop2=c("EUR","NAT","NAT"))

gla

Optional 2 x K numeric matrix (rows "p"/"q", columns matching names(pure_codes)) of per-arm global local ancestry proportions, as produced by ancestry_split internally. When NULL (default), GLA shrinkage is disabled and behavior matches the original p[k] estimator exactly, including its 0.5/0.5 singleton fallback.

arm_id

Optional integer vector (length = nrow(gt_genotype)) giving each variant's row index into gla (0-based). Required together with gla.

Value

Named list of K numeric matrices (variants x samples), one per population in pure_codes, in the same order. Each element is named after the corresponding entry in pure_codes.

Examples

# 3-population example: AFR=3, EUR=1, NAT=4; mixed codes 2/5/6
set.seed(1)
gt  <- matrix(sample(0:2, 20, replace = TRUE, prob = c(.8,.15,.05)),
              nrow = 4, ncol = 5)
# ancestry codes: 3=AFR/AFR, 1=EUR/EUR, 4=NAT/NAT,
#                 2=AFR/EUR, 5=AFR/NAT, 6=EUR/NAT
anc <- matrix(sample(c(1L,2L,3L,4L,5L,6L), 20, replace = TRUE),
              nrow = 4, ncol = 5)
pure  <- c(AFR = 3L, EUR = 1L, NAT = 4L)
mixed <- data.frame(code = c(2L, 5L, 6L),
                    pop1 = c("AFR", "AFR", "EUR"),
                    pop2 = c("EUR", "NAT", "NAT"))
out <- split_diploid_multi(gt, anc, pure, mixed)
names(out)   # "AFR" "EUR" "NAT"
#> [1] "AFR" "EUR" "NAT"

Split unphased genotype matrix by ancestry for K populations

Arguments

Value

See also

Examples