Convert R plotting commands into loop for many files - r

Question

Currently I have been using R to read in a table and plot some of the data which I save as a png file. Now I have 100 files and would like this process to be automated rather than manually changing the path 100 times.
Additionally I would like to join the 100 files into one table in R that I can subsequently analyse. The join would be in the format of dplyr's bind_rows as all files have the same column headers. I've done this for when I have two tables in R but now when I am using a loop to read files in sequentially. What would be the best way to do this in R? Thanks in advance for any suggestions or help.
my_data <- read.table('/path/to/data/results/sample_1.txt', header = TRUE, sep = "\t")
ggplot(my data, aes(x=alt_freq)) + geom_histogram(color="black", fill="white", bins = 20) + xlim(c(0,1))
ggsave("/path/to/plots/sample_1.png", plot = last_plot(),width = 16, height = 9)
#append table to one large table in the format of dplyr::bind_rows(y, z)
Input files are all named with the same naming convention:
sample_1.txt
sample_2.txt
sample_3.txt
The files look like:
sample_name position alt_freq ref_freq sample_1_counts
sample 1 10 0.5 0.5 2
sample 1 20 0.25 0.75 4
All txt files are in the same directory and all txt files are of interest.

Answer 1

First collect the complete path of the files of interest
library(ggplot2)
all_files <- list.files("/path/to/data/results", pattern = "sample_\\d+\\.txt$",
full.names = TRUE)
Then create a function to apply to each file
new_fun <- function(path_of_file) {
my_data <- read.table(path_of_file, header = TRUE)
ggplot(my_data, aes(x=alt_freq)) +
geom_histogram(color="black", fill="white", bins = 20) + xlim(c(0,1))
ggsave(paste0(dirname(path_of_file), "/", sub("txt$", "png",
basename(path_of_file))), plot = last_plot(),width = 16, height = 9)
}
We use paste0 to create path to save the plot dynamically by getting the directory name and replacing the ending txt with png.
Then use lapply/map/for loop to apply new_fun to each file
lapply(all_files, new_fun)
To combine all the files into one dataframe we can do
combined_data <- do.call(rbind, lapply(all_files, read.table, header = TRUE))
If the header is different for one column we can change the column name for that particular column and then rbind. So for example, if the header information for column 1 is different, we can do
combined_data <- do.call(rbind, lapply(all_files, function(x) {
df <- read.table(x, header = TRUE)
names(df)[1] <- "new_header"
df$filename <- basename(x)
df
}))

Answer 2

I would do something like the following.
Change these to their real values.
in_dir <- '/path/to/data/results'
out_dir <- '/path/to/plots'
Now the plots and binding the tables.
library(ggplot2)
old_dir <- getwd()
setwd(in_dir)
flnames <- list.files(pattern = '^sample_[[:digit:]]+\\.txt$')
data_list <- lapply(flnames, read.table, header = TRUE, sep = '\t')
lapply(seq_along(data_list), function(i){
ggplot(data_list[[i]], aes(x = alt_freq)) +
geom_histogram(color = "black", fill = "white", bins = 20) +
xlim(c(0, 1))
f <- sub('txt$', 'png', flname[i])
outfile <- paste(out_dir, f, sep = '/')
ggsave(outfile, plot = last_plot(),width = 16, height = 9)
})
data_all <- dplyr::bind_rows(data_list)
Final cleanup.
setwd(old_dir)
## NOT RUN
#rm(data_list)