R/pairwise_distances.R

In sams25/rcombinator_old: Tools to Simulate the Evolution of Genetic Sequences

#' Gets pairwise distances from the sequence data of a Point mutation simulation
#' @param seq_data A data frame that contains a list of sequences for each time
#' @return A data frame with the pairwise distances between the sequences for each time step
#' @export
pairwise_distances <- function(seq_data)
{
  num_jumps <- length(seq_data$time)
  num_sequences <- length(seq_data$sequences[[num_jumps]])
  num_seq_pairs <- (num_sequences * (num_sequences-1))/2
  size <- num_seq_pairs * num_jumps

  pair_data <- data.frame(time = rep(NA_integer_, size),
                          seq1 = rep(NA_integer_, size),
                          seq2 = rep(NA_integer_, size),
                          seq_dist = rep(NA, size))

  assign_row <- function(time_index, seq_data, num_seq_pairs)
  {
    new_rows <- dist_to_data_frame(stringdist_ape(seq_data$sequences[[time_index]]))

    assign_row_inner <- function(seq_index, new_rows, time_index)
    {
      row_index <- (time_index-1) * num_seq_pairs + seq_index
      new_row <- list(time_index,
                      new_rows$seq1[seq_index], new_rows$seq2[seq_index],
                      new_rows$dist_1_2[seq_index])
      pair_data[row_index, ] <<- new_row
    }
    purrr::map(1:num_seq_pairs, assign_row_inner, new_rows, time_index)
  }
  purrr::map(1:num_jumps, assign_row, seq_data, num_seq_pairs)

  return(pair_data)
}