Detects where adding more genes yields less than a threshold improvement
in the metric of interest (e.g. S_AB). Uses a smoothed first difference.
Usage
find_saturation_point(x, y, threshold = 0.01)
Arguments
- x
Numeric vector of x-values (e.g. N).
- y
Numeric vector of y-values (e.g. S_AB).
- threshold
Minimum absolute change required to continue adding
genes. Default 0.01.
Value
An integer index into x where saturation is reached.
Examples
x <- seq(10, 100, by = 10)
y <- c(0.5, 0.3, 0.15, 0.08, 0.051, 0.049, 0.048, 0.047, 0.046, 0.046)
idx <- find_saturation_point(x, y, threshold = 0.01)
x[idx]
#> [1] 60
