Skip to contents

K-nearest-neighbors by TCRdist with optional group masking

Source: R/wrappers_neighbors.R
tcrdist_knn.Rd

For each of the N input TCRs, finds the K nearest neighbors by TCRdist distance. TCRs sharing the same agroups or bgroups value are excluded from each other's neighborhood (same-group masking). When agroups and bgroups are NULL (default), every TCR has its own unique group so no masking occurs.

Usage

tcrdist_knn(
  tcrs,
  organism,
  K,
  agroups = NULL,
  bgroups = NULL,
  sort_nbrs = TRUE,
  components = "all",
  weight_cdr3 = WEIGHT_CDR3_REGION,
  gap_penalty_cdr3 = GAP_PENALTY_CDR3_REGION
)

Arguments

tcrs

A data.frame with at least the following columns:

va

Character. Alpha-chain V-gene allele.

cdr3a

Character. Alpha-chain CDR3 amino acid sequence.

vb

Character. Beta-chain V-gene allele.

cdr3b

Character. Beta-chain CDR3 amino acid sequence.

organism

Character string. Organism key, e.g. "human".

K

Integer. Number of nearest neighbors to return per TCR. Must satisfy 1 <= K <= N - 1.

agroups

Integer vector of length N, or NULL. Alpha-chain group assignments. TCRs with the same value are masked from each other. NULL assigns each TCR a unique group (no masking).

bgroups

Integer vector of length N, or NULL. Beta-chain group assignments. Same semantics as agroups.

sort_nbrs

Logical. If TRUE (default), sort each row's K neighbors by ascending distance.

components

Character. Which distance components to include. See tcrdist_matrix for details. Default "all".

weight_cdr3

Integer. CDR3 distance weight. Defaults to WEIGHT_CDR3_REGION (3L).

gap_penalty_cdr3

Integer. CDR3 gap penalty. Defaults to GAP_PENALTY_CDR3_REGION (12L).

Value

A list with two elements:

knn_indices

Integer matrix (N x K). 1-based row indices of the K nearest neighbors for each TCR.

knn_distances

Numeric matrix (N x K). Corresponding TCRdist distances.

See also

tcrdist_radius_neighbors, knn_from_matrix, knn_from_pca

Examples

# \donttest{
tcrs <- data.frame(
  va    = c("TRAV1-1*01", "TRAV1-1*01", "TRAV12-2*01"),
  cdr3a = c("CAVRDSSYKLIF", "CAVRDSSYKLIF", "CAVSANSGTYF"),
  vb    = c("TRBV19*01", "TRBV19*01", "TRBV20-1*01"),
  cdr3b = c("CASSIRSSYEQYF", "CASSIRSYEQYF", "CSARDRTGNTIYF"),
  stringsAsFactors = FALSE
)
knn <- tcrdist_knn(tcrs, "human", K = 1L)
# }