You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
One of the methods for selecting OTU's from the ranked BC matrix is the 'elbow' method.
According to the Shade and Stopnisiek paper: The method to find the ‘elbow’ of a graph is based on the first-order difference (from numerical differentiation). In this method, a score is assigned to each cutoff point by separating the curve into two parts and calculating the difference in the average rates of change for both of these parts. Then, the elbow point is selected to be the cutoff point that maximizes this score.
In the ExtractCore.R function, this first-order differentiation is supposedly calculated in
I'm unsure what the 'pos' value is supposed to be. As well as the 1:elbow used to extract core_otus. The function DOES run when method 'elbow' is specified... but it does not run line-by-line because 'pos' is an undefined variable.
We aren't currently using the 'elbow' method, and it is supposedly more stringent than the 'increase' method. But I do not think that this function is likely running the way it is intended.
The text was updated successfully, but these errors were encountered:
One of the methods for selecting OTU's from the ranked BC matrix is the 'elbow' method.
According to the Shade and Stopnisiek paper: The method to find the ‘elbow’ of a graph is based on the first-order difference (from numerical differentiation). In this method, a score is assigned to each cutoff point by separating the curve into two parts and calculating the difference in the average rates of change for both of these parts. Then, the elbow point is selected to be the cutoff point that maximizes this score.
In the ExtractCore.R function, this first-order differentiation is supposedly calculated in
function(pos){ left <- (BC_ranked[pos, 2] - BC_ranked[1, 2]) / pos right <- (BC_ranked[nrow(BC_ranked), 2] - BC_ranked[pos, 2]) / (nrow(BC_ranked) - pos) core_otus <- otu_ranked$otu[1:elbow]}
I'm unsure what the 'pos' value is supposed to be. As well as the 1:elbow used to extract core_otus. The function DOES run when method 'elbow' is specified... but it does not run line-by-line because 'pos' is an undefined variable.
We aren't currently using the 'elbow' method, and it is supposedly more stringent than the 'increase' method. But I do not think that this function is likely running the way it is intended.
The text was updated successfully, but these errors were encountered: