mirror of
https://github.com/johrpan/ubigen.git
synced 2025-10-26 19:57:24 +01:00
60 lines
2.4 KiB
R
60 lines
2.4 KiB
R
# This scripts reads the input data (See input.R) and performs various
|
|
# computations on it in order to later use the results for computating scores
|
|
# for ubuiquitously expressed genes.
|
|
|
|
library(data.table)
|
|
library(here)
|
|
|
|
i_am("scripts/input.R")
|
|
|
|
data <- fread(here("scripts", "input", "data_long.csv"))
|
|
|
|
data[, `:=`(
|
|
expression_median = median(expression),
|
|
expression_95 = quantile(expression, probs = 0.95)
|
|
), by = sample]
|
|
|
|
# Transform the expression logarithmically. The samples that don't express a
|
|
# gene at all will be left out intentionally.
|
|
data[expression > 0, expression_log := log2(expression)]
|
|
|
|
results <- data[, .(
|
|
median_expression = median(expression[expression > 0]),
|
|
iqr_expression = IQR(expression[expression > 0]),
|
|
mean_expression = mean(expression[expression > 0]),
|
|
sd_expression = sd(expression[expression > 0]),
|
|
median_expression_normalized = median(expression_log, na.rm = TRUE),
|
|
iqr_expression_normalized = IQR(expression_log, na.rm = TRUE),
|
|
mean_expression_normalized = mean(expression_log, na.rm = TRUE),
|
|
sd_expression_normalized = sd(expression_log, na.rm = TRUE),
|
|
above_zero = mean(expression > 0.0),
|
|
above_threshold = mean(expression > 50.0),
|
|
above_median = mean(expression > expression_median),
|
|
above_95 = mean(expression > expression_95)
|
|
), by = "gene"]
|
|
|
|
# Normalize values to the range of 0.0 to 1.0.
|
|
results[, `:=`(
|
|
median_expression_normalized =
|
|
(median_expression_normalized -
|
|
min(median_expression_normalized, na.rm = TRUE)) /
|
|
(max(median_expression_normalized, na.rm = TRUE) -
|
|
min(median_expression_normalized, na.rm = TRUE)),
|
|
iqr_expression_normalized =
|
|
(iqr_expression_normalized -
|
|
min(iqr_expression_normalized, na.rm = TRUE)) /
|
|
(max(iqr_expression_normalized, na.rm = TRUE) -
|
|
min(iqr_expression_normalized, na.rm = TRUE)),
|
|
mean_expression_normalized =
|
|
(mean_expression_normalized -
|
|
min(mean_expression_normalized, na.rm = TRUE)) /
|
|
(max(mean_expression_normalized, na.rm = TRUE) -
|
|
min(mean_expression_normalized, na.rm = TRUE)),
|
|
sd_expression_normalized =
|
|
(sd_expression_normalized -
|
|
min(sd_expression_normalized, na.rm = TRUE)) /
|
|
(max(sd_expression_normalized, na.rm = TRUE) -
|
|
min(sd_expression_normalized, na.rm = TRUE))
|
|
)]
|
|
|
|
fwrite(results, file = here("scripts", "output", "results.csv"))
|