In philliplab/MotifBinner: Bin reads by Motif

library(Biostrings)
options(scipen=999)
bbn_dat <- report_dat$bbn_dat
attach(bbn_dat)

A Basic report on the unprocessed bins produced

Sys.time()

The number of bins

print(length(bin_seqs))

The sizes of the bins

bin_sizes <- unlist(lapply(bin_seqs, length))
table(bin_sizes)
hist(bin_sizes)

log log plot of the bin sizes

loglog <- data.frame(bin_size = as.numeric(names(table(bin_sizes))),
                     num_bins = as.numeric(table(bin_sizes)))
plot(num_bins ~ bin_size, data = loglog)
plot(log10(num_bins) ~ log10(bin_size), data = loglog)

Detailed investigations of the distances in the bins

comp_bins <- NULL
bin_dists <- list()
for (i in seq_along(bin_seqs)){
  the_bin <- bin_seqs[[i]]
  if (length(the_bin) > 100){
    the_bin <- the_bin[sample(1:length(the_bin), 100)]
  }
  dmat <- stringDist(the_bin)  
  bin_name <- strsplit(names(the_bin)[1], '_')[[1]][1]
  bin_dists[[bin_name]] <- dmat
  max_occ <- max(table(the_bin))
  num_max <- sum(table(the_bin) == max_occ)
  comp_bins <- rbind(comp_bins,
                     data.frame(bin_id = i,
                                bin_name = bin_name,
                                bin_size = length(bin_seqs[[i]]),
                                working_size = length(the_bin),
                                min_dist = min(dmat),
                                max_dist = max(dmat),
                                mean_dist = mean(dmat),
                                min_seq_len = min(width(bin_seqs[[i]])),
                                max_seq_len = max(width(bin_seqs[[i]])),
                                mean_seq_len = mean(width(bin_seqs[[i]])),
                                max_occ = max_occ,
                                num_max = num_max
                                ))
}

Sequence lengths

hist(comp_bins$min_seq_len)
hist(comp_bins$mean_seq_len)
hist(comp_bins$max_seq_len)

Bins of size 2

c_bins <- subset(comp_bins, bin_size == 2)
hist(c_bins$mean_dist)
kable(data.frame(bin_dist = as.numeric(names(table(c_bins$mean_dist))),
                 num_bins = as.numeric(table(c_bins$mean_dist))))

Bins of size 3

c_bins <- subset(comp_bins, bin_size == 3)
hist(c_bins$min_dist)
kable(data.frame(min_dist = as.numeric(names(table(c_bins$min_dist))),
                 num_bins = as.numeric(table(c_bins$min_dist))))
hist(c_bins$max_dist)
kable(data.frame(max_dist = as.numeric(names(table(c_bins$max_dist))),
                 num_bins = as.numeric(table(c_bins$max_dist))))
plot(c_bins$min_dist ~ c_bins$max_dist)

Bins of size 4 and above

c_bins <- subset(comp_bins, bin_size > 3)
hist(c_bins$bin_size)

hist(c_bins$min_dist)
kable(data.frame(min_dist = as.numeric(names(table(c_bins$min_dist))),
                 num_bins = as.numeric(table(c_bins$min_dist))))

hist(c_bins$mean_dist)
kable(data.frame(mean_dist = as.numeric(names(table(round(c_bins$mean_dist,0)))),
                 num_bins = as.numeric(table(round(c_bins$mean_dist,0)))))

hist(c_bins$max_dist)
kable(data.frame(max_dist = as.numeric(names(table(c_bins$max_dist))),
                 num_bins = as.numeric(table(c_bins$max_dist))))
plot(c_bins$min_dist ~ c_bins$max_dist)
hist(c_bins$max_occ)
kable(data.frame(max_occ = as.numeric(names(table(c_bins$max_occ))),
                 num_bins = as.numeric(table(c_bins$max_occ))))