Deduplicate alignment ribbons

Removes fully duplicated, coordinate-near-duplicated or highly overlapping alignment blocks within each (query, subject) pair and keeps the best representative per keep.

Usage

deduplicate_ggchord_ribbons(
  ribbon_data,
  tolerance = 0,
  by = c("exact", "coordinates", "overlap"),
  keep = c("best_pident", "longest", "first"),
  min_reciprocal_overlap = 0.9
)

Arguments

ribbon_data

data.frame in ribbon_data format.

tolerance

Numeric, default 0. Maximum absolute difference (in bp) allowed on each of qstart/qend/sstart/send for two rows to count as coordinate duplicates (used with by = "coordinates").

by

Character, default "exact". "exact" removes rows with identical pair and coordinates; "coordinates" additionally treats rows within tolerance bp as duplicates; "overlap" treats blocks whose reciprocal overlap is at least min_reciprocal_overlap on both sequences as duplicates.

keep

Character, default "best_pident". Which representative to keep: "best_pident" (highest pident), "longest" (longest length) or "first" (first occurrence in the input).

min_reciprocal_overlap

Numeric (0-1), default 0.9. Reciprocal overlap threshold used with by = "overlap".

Value

A list with data (deduplicated data frame with source_rows attribute) and report (n_input/n_kept/n_removed plus a data.frame of removed rows with row, duplicate_of and reason).

Examples

library(ggchord)
data(ribbon_data_example)
dup <- rbind(ribbon_data_example, ribbon_data_example[1, ])
out <- deduplicate_ggchord_ribbons(dup, by = "exact")
out$report
#> $n_input
#> [1] 32
#> 
#> $n_kept
#> [1] 31
#> 
#> $n_removed
#> [1] 1
#> 
#> $removed
#>   row duplicate_of          reason
#> 1  32            1 duplicate_exact
#>