Merge adjacent or overlapping alignment blocks of the same sequence pair

Merges alignment blocks that belong to the same (query, subject) pair, are adjacent (gap <= max_gap) or overlapping on both sequences, and are compatible. The merged pident is weighted by alignment length. Merging is deliberately conservative: blocks whose merged query and subject spans would be inconsistent (unequal), whose pident differs by more than min_pident_difference, or whose orientation differs (when require_same_orientation) are left unmerged.

Usage

merge_ggchord_ribbons(
  ribbon_data,
  max_gap = 0,
  min_pident_difference = 0,
  require_same_orientation = TRUE,
  group_by = c("qaccver", "saccver")
)

Arguments

ribbon_data

data.frame in ribbon_data format.

max_gap

Numeric, default 0. Maximum gap (in bp) allowed between two blocks on both sequences for them to be merged.

min_pident_difference

Numeric, default 0. When > 0, two blocks are only merged when their pident difference is <= this value.

require_same_orientation

Logical, default TRUE. Only merge blocks whose query/subject direction is the same (both ascending or both descending, i.e. collinear or both inverted in the same way).

group_by

Character vector, default c("qaccver", "saccver"). Columns used to identify the same sequence pair.

Value

A list with data (merged data frame with source_rows attribute) and report (data.frame with output_row, from_rows and n_merged for every output row).

Examples

library(ggchord)
# Two adjacent, collinear blocks of the same pair -> one merged block
rb <- data.frame(
  qaccver = c("A", "A"), saccver = c("B", "B"),
  length = c(100, 100), pident = c(95, 97),
  qstart = c(1, 101), qend = c(100, 200),
  sstart = c(501, 601), send = c(600, 700)
)
out <- merge_ggchord_ribbons(rb, max_gap = 0)
out$data
#>   qaccver saccver length pident qstart qend sstart send
#> 1       A       B    200     96      1  200    501  700
out$report
#>   output_row from_rows n_merged
#> 1          1       1,2        2