Pass arguments to the R Phyloseq subset_taxa wrapper - r

Question

I'll explain the end goal, and what I'm trying as a test first. (Because I'm likely going about it the wrong way.)
I am using the phyloseq package to visualize microbiome data. I want to "automate" it to an extent by having users choose levels of analysis and have my script generate the visualizations without someone hand typing in each combination.
The issue is passing variables into the subset function. I get these errors primarily (depending on what combinations of paste0, eval, parse, as.logical, expression, noquote....etc that i've tried):
Error in subset.data.frame(oldDF, ...) : 'subset' must be logical
Error in dimnames(x) <- dn :
length of 'dimnames' [1] not equal to array extent
A user would set the levels of analysis. So lets say for now there are two levels, and selecting the second level automatically means you want the first level as well. (I haven't worked on that part yet, but I wanted to explain it upfront.
#Set lineage level
lin_level <- 1
lin_list <- c("k__Kingdom", "p__Phylum","c__Class", "o__Order","f__Family")
lin_select <- lin_list[lin_level]
sub_lin <- lin_list[(lin_level +1)]
#Kingdom
king_list <- "k__Bacteria"
#set Phylum list
if (lin_select == "p__Phylum"){
phylum_list <- c("p__Firmicutes","p__Proteobacteria","p__Bacteroidetes","p__Actinobacteria","p__Tenericutes")
}
subgroup <- "All"
From here, the script would ultimately get to the graphing section. If lin_level is set to 1, it would look like this:
FIXED
gphic = subset_taxa(physeq1, Kingdom=="k__Bacteria")
title = paste0(subgroup," ", "Bacteria-only")
plot_bar(gpsfb, "Phylum", "Abundance", "Phylum",
title=title, facet_grid="Type~.")
AUTOMATED
gphic = subset_taxa(physeq1, (substring(lin_select,4)) == king_list)
title = paste0(subgroup," ", (substring(king_list,4)),"-only")
plot_bar(gpsfb, (substring(sub_lin,4)), "Abundance", (substring(sub_lin,4)),
title=title, facet_grid="Type~.")
But, trying to pass (substring(lin_select,4)) == king_list as an argument results in errors.
I've searched through the various threads on this issue, but haven't been able to get the different answers to work. Ultimately I need to run the graphing section once for Kingdom, and then again each time for each item in the Phylum list. But before i can get there, I need to be able to pass the arguments into the subset function.
Things I've tried:
test <- paste0(substring(lin_select,4),"==","\"","p__Bacteroidetes","\"")
noquote(test)
[1] Phylum=="p__Bacteroidetes"
gphic = subset_taxa(physeq1, noquote(test))
Error in subset.data.frame(oldDF, ...) : 'subset' must be logical
gphic = subset_taxa(physeq1, paste0(substring(lin_select,4),"==","\"","p__Bacteroidetes","\""))
Error in subset.data.frame(oldDF, ...) : 'subset' must be logical
gphic = subset_taxa(physeq1, as.logical(test))
Error in dimnames(x) <- dn :
length of 'dimnames' [1] not equal to array extent
as.logical(noquote(test))
[1] NA
gphic = subset_taxa(physeq1, as.logical(noquote(test)))
Error in dimnames(x) <- dn :
length of 'dimnames' [1] not equal to array extent
noquote(test)
[1] Phylum=="p__Bacteroidetes"
as.logical(noquote(test))
[1] NA
as.logical(as.character(noquote(test)))
[1] NA
test2 <- eval(parse(text= test))
Error in eval(parse(text = test)) : object 'Phylum' not found
test2 <- eval(test)
gphic = subset_taxa(physeq1, as.logical(test2))
Error in dimnames(x) <- dn :
length of 'dimnames' [1] not equal to array extent
as.logical(test2)
[1] NA
And a lot of other permutations trying to sub in different things, but you get the idea.

Answer 1

gphic = subset_taxa(physeq1, eval(as.name(level_tax)) == king_list)
Here , level_tax is the variable in a loop. Say level_tax = "Order", then we convert the string "Order" into variable name by as.name(level_tax) or as.symbol(level_tax). Then we use eval(), which takes an expression and evaluates in the specified environment