Micropower with weighted unifrac

I'm trying to run Micropower using weighted unifrac distance matrices following the outline found here: https://gist.githubusercontent.com/brendankelly/6673e8596d3cde3fac7d493d7747aa80/raw/59ee6a6be3db8bd751817cdd7c31c7bd12d4d5f8/micropower_simulation_example.R

For the following command,

sd_wu_df <- data.frame(otu_number=round(as.numeric(gsub(".txt","",gsub("otu","",sapply(strsplit(names(sd_wu),"_"),FUN=function(x){x[[4]]})), fixed = TRUE))),sd=sapply(sd_wu, FUN=function(x){sd(as.dist(x))}))

I get this: Error in x[[4]] : subscript out of bounds

The script will only run if I sub 1 in for 4 as follows:

sd_wu_df <- data.frame(otu_number=round(as.numeric(gsub(".txt","",gsub("otu","",sapply(strsplit(names(sd_wu),"_"),FUN=function(x){x[[1]]})), fixed = TRUE))),sd=sapply(sd_wu, FUN=function(x){sd(as.dist(x))}))

But then I get this:
Warning message:
In data.frame(otu_number = round(as.numeric(gsub(".txt", "", gsub("otu", :
NAs introduced by coercion

That causes my table to have NA for otu_number, meaning I can't do the next steps in the script. Screen shot below. What is the problem with the script?

Figured it out if anyone else runs into the same problem. Since my tables were output as .tsv files, I just had to switch .txt to .tsv. Still had to sub in [1] for [4]. Then it ran fine.