johrpan/geposanui
1
0
Fork 0
mirror of https://github.com/johrpan/geposanui.git synced 2025-10-26 19:27:24 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions

Include all species

Browse source
This commit is contained in:
Elias Projahn 2021-08-25 12:00:22 +02:00
parent abab22a907
commit 129793e055
1 changed files with 24 additions and 21 deletions
Download patch file Download diff file Expand all files Collapse all files

45
data.R
View file

@ -1,4 +1,5 @@
library(data.table) library(data.table)
library(rlog)
#' Merge genome data from files in `path` into `data.table`s. #' Merge genome data from files in `path` into `data.table`s.
#' #'
@ -26,35 +27,37 @@ load_data <- function(path) {
# Each file will contain data on one species. # Each file will contain data on one species.
file_names <- list.files(paste(path, "genomes", sep = "/")) file_names <- list.files(paste(path, "genomes", sep = "/"))
n_species <- length(file_names)
for (file_name in file_names) { for (i in seq_along(file_names)) {
file_name <- file_names[i]
species_id <- strsplit(file_name, split = ".", fixed = TRUE)[[1]][1] species_id <- strsplit(file_name, split = ".", fixed = TRUE)[[1]][1]
species_path <- paste(path, "genomes", file_name, sep = "/")
# Only continue for replicatively aging species. log_info(sprintf(
# TODO: Which other species should be included? "Reading species %i/%i (%s)", i, n_species, species_id
if (original_species[id == species_id, group] == "replicative") { ))
species_path <- paste(path, "genomes", file_name, sep = "/")
species_distances <- fread(species_path)
# Compute the median distance across all genes of this species and species_distances <- fread(species_path)
# add it to the species table along other static data.
species <- rbindlist(list(species, data.table(
id = species_id,
label = original_species[id == species_id, label],
median_distance = median(species_distances[, dist])
)))
species_distances <- data.table( # Compute the median distance across all genes of this species and
species = species_id, # add it to the species table along other static data.
gene = species_distances[, geneid], species <- rbindlist(list(species, data.table(
distance = species_distances[, dist] id = species_id,
) label = original_species[id == species_id, label],
median_distance = median(species_distances[, dist])
)))
distances <- rbindlist(list(distances, species_distances)) species_distances <- data.table(
} species = species_id,
gene = species_distances[, geneid],
distance = species_distances[, dist]
)
distances <- rbindlist(list(distances, species_distances))
} }
# Order species by there median distance. # Order species by their median distance.
setorder(species, median_distance) setorder(species, median_distance)
list( list(