In dmontemayor/Rcricvol: Data Processing Tools in R for cricvol project

Objective: Pre-process all untargeted data (log transformation and z score)

# load annotated ions
original_untar <- as.data.frame(t(read.csv("../data/CRIC_clinical_creatinine_normalized_data.csv")))
ann <- read.csv("../data/CRIC_annotation_1mD_neg.csv")
untar <- original_untar[,names(original_untar) %in% paste0("V", ann$ion)]

log_untar <- log(untar)

#Screen ions to get to about ~5000. use coefficient of variation (used to make selected zscored ions)
#test <- data.frame(apply(log_original_untar, MARGIN = 2, function(x){sd(x)/mean(x)}))
#names(test) <- c("cov")

# ions with coefficient of variation greater than 1
#cutoff <- quantile(test$cov, .66)
#ions <- log_original_untar[which(test$cov > cutoff)]

Z score ions

zscore_untar <- as.data.frame(apply(log_untar, MARGIN = 2, function(x) {
  (x - mean(x, na.rm = TRUE)) / var(x, na.rm = TRUE)
}))

Add in patient ID and group and visit

untarmet <- read.csv("../data/CRIC_clinical_creatinine_normalized_samplemetadata.csv", header = FALSE)
partitioning <- read.csv("../data/partitioning.csv")
names(partitioning)[1] <- "patientid"

zscore_untar$patientid <- substring(untarmet$V2, 1, 7)
zscore_untar$visit <- substring(untarmet$V2, 9, 15)
zscore_untar <- merge(zscore_untar, partitioning[c(1,2)], by = "patientid") # combine groupings and patient ids in order

zscore_untar <- zscore_untar[c(1,(length(zscore_untar)-1), length(zscore_untar), 2:(length(zscore_untar)-2))]
write.csv(zscore_untar, "../data/zscore_untargeted_annotatedions.csv", row.names = FALSE, col.names = colnames(zscore_untar))

Example:

Before:

hist(original_untar[,1])

After:

hist(zscore_untar[,4])

dmontemayor/Rcricvol documentation built on Sept. 9, 2021, 9:12 a.m.