Read 10X Genomics VDJ contig annotations

Reads a 10X Genomics cellranger VDJ output file (typically filtered_contig_annotations.csv) and returns a paired alpha-beta TCR data frame suitable for TCRrep. Each row in the 10X format represents a single contig; this function filters, deduplicates, and pairs alpha and beta chains by cell barcode.

Usage

read_10x(
  file,
  pair_by = "barcode",
  productive_only = TRUE,
  normalize_genes = TRUE
)

Arguments

file: Character string. Path to a 10X VDJ CSV file (e.g. filtered_contig_annotations.csv).
pair_by: Character string. Column name used to pair TRA and TRB contigs from the same cell. Defaults to "barcode".
productive_only: Logical. If TRUE (default), only productive contigs are retained (productive == "True").
normalize_genes: Logical. If TRUE (default), gene names lacking an allele suffix (e.g. "TRAV1-1") are appended with *01.

Value

A data.frame with columns barcode, va, ja, cdr3a, vb, jb, cdr3b. Only cells with both a TRA and TRB contig are included (inner join).

Details

When a cell has multiple contigs for the same chain, the contig with the highest UMI count is retained (if the umis column is present); otherwise the first contig is kept.

Examples

# \donttest{
# df <- read_10x("filtered_contig_annotations.csv")
# obj <- TCRrep(df, organism = "human")
# }