Filter Dorado summary for reads fulfilling ninetails quality criteria

Takes a Dorado summary file (from ONT Dorado basecaller) or a data frame containing equivalent summary information, and filters out reads that do not meet ninetails quality and alignment criteria.

Usage

filter_dorado_summary(dorado_summary)

Arguments

dorado_summary

Character string or data frame. Path to a tab-delimited Dorado summary file, or a data frame containing equivalent summary information. Must contain at least the columns read_id and poly_tail_length.

Value

A filtered tibble or data frame containing only reads that meet the filtering criteria.

Details

Reads must meet all four criteria to pass the filter:

1. Mapped: alignment_direction is "+" or "-", not "*"

2. High mapping quality: alignment_mapq > 0

3. Valid coordinates: poly_tail_start != 0 (in DRS, the adapter passes through the pore first, so a start position of 0 indicates a likely artifact)

4. Sufficient length: poly_tail_length >= 10 nt (shorter tails cannot be reliably processed by the CNN)

See also

preprocess_inputs and preprocess_inputs_cdna where this function is called during pipeline initialization, process_dorado_summary for splitting large summary files

Examples

if (FALSE) { # \dontrun{

# From file
filtered <- filter_dorado_summary("dorado_summary.txt")

# From data frame
df <- data.frame(
  read_id = c("read1", "read2"),
  alignment_direction = c("+", "*"),
  alignment_mapq = c(60, 0),
  poly_tail_start = c(100, 0),
  poly_tail_length = c(20, 5)
)
filtered <- filter_dorado_summary(df)

} # }