trim_cor: Calculates correlation of shifted vectors
In Swarchal/TISS: Titration Invariant Similarity Score

Description Usage Arguments Value Examples

This function is for calculating the correlation between two vectors x and y. The vectors are shifted n times by p elements. This relates to shifted n number of concentrations in either direction, each concentration shift being a multiple of the number of features.

1	trim_cor(x, y, n = NULL, p = NULL, metadata = NULL)

`x`	vector
`y`	vector
`n`	integer. The number of titrations shifts in either directions. i.e `n = 2` will shift 2 titrations to the left, and 2 to the right. If n is not directly given, then it will be calculated from by half the number of concentrations in the metadata (to the nearest integer).
`p`	interger. The number of features. `n` will always be a multiple of `p`. This is so identical features are always aligned despite titration shifts.
`metadata`	optional metadata object produced from `construct_metadata`. Used to extract values for n and p if none are given.

cor_measurements vector of correlations for each shift

# create dummy dataset
l <- list()
for (i in 1:100) l[[i]] <- rnorm(100)
df <- as.data.frame(l)
names(df) <- paste("col", 1:100, sep = "_")
# order values (more obvious correlation shifts)
df <- apply(df, 2, sort)
out <- trim_cor(df[,1], df[,2], n = 5, p = 5)
plot(out, type = "l")

# plot all correlations
out <- trim_cor(df[,1], df[,2], 20, 3)
plot(out,
     type = "l",
     xaxt =  "n",
     ylab = "correlation",
     main = "col_1 vs all correlation",
     sub = "20 shifts, 3 elements each")
axis(1, at = 1:41,
     labels = names(out))
for (i in 3:ncol(df)){
 lines(trim_cor(df[,1], df[,i], 20, 3), col = "gray50")
}